JUAN DIEGO URRJAGO SUAREZ

Les réponses comportementales de P oursin niger face aux étoiles de mer prédatrices gelatinosus et helianthus

Mémoire présenté à la Faculté des études supérieures de l'Université Laval dans le cadre du programme de maîtrise en biologie pour l'obtention du grade de Maître es Sciences (M.Sc.)

DEPARTEMENT DE BIOLOGIE FACULTÉ DES SCIENCES ET DE GÉNIE UNIVERSITÉ LAVAL QUÉBEC

2010

© Juan Diego Urriago, 2010 Résumé J'ai mené des expériences afin d'étudier les réponses de l'oursin Tetrapygus niger à la prédation des étoiles de mer. L'oursin était capable de différencier les étoiles de mer prédatrices des non-prédatrices mais également de distinguer différents niveaux de risques associés aux étoiles de mer, et Meyenaster gelatinosus. Les oursins soumis à un haut niveau de risque hérissaient rapidement leurs épines, puis étendaient leurs pieds ambulacraires pour fuir le prédateur. L'oursin associait un risque plus grand à M. gelatinosus. J'ai également démontré l'existence de la chimiodétection à distance des prédateurs. La micro-distribution des oursins sur les surfaces élevées semble représenter une stratégie pour limiter la prédation par les étoiles de mer. Le nombre de celles-ci étant plus réduit sur ces surfaces élevées, leur capacité à capturer les oursins est moindre et les oursins peuvent se détacher pour éviter d'être mangé. Enfin, des expériences avec entravement indiquent que le taux de survie est plus élevé pour les animaux situés sur les surfaces élevées. 11

Abstract I conducted field experiments to examine responses of the Tetrapygus niger to sea star predators. The urchin distinguished between predatory and non-predatory sea stars and also recognized different levels of risk associated with the predatory sea stars, Heliaster helianthus and Meyenaster gelatinosus. Urchins under high prédation risk rapidly raised their spines, and then extended podia and fled. The urchin associates the strongest risk with M. gelatinosus. I further demonstrated distance chemodetection of predators. The urchin's micro-distribution on elevated surfaces appears to represent a strategy for limiting prédation by sea stars because on elevated surfaces there are fewer sea stars, the ability of sea stars to capture urchins is reduced, and the urchins can detach to avoid being eaten. Finally, tethering experiments indicate that survival rate is greater on elevated surfaces. Ill

Avant-propos Ce mémoire comporte quatre chapitres : une introduction générale, deux chapitres dans lesquels les résultats obtenus au cours de mon étude sont présentés et discutés (corps du mémoire) et une conclusion générale. Les chapitres qui constituent le corps du mémoire ont été rédigés en anglais sous la forme d'articles scientifiques. J'ai été l'instigateur et le réalisateur de chacune de ces études. Les Drs. John H. Himmelman (directeur du mémoire) et Carlos F. Gaymer (codirecteur) ont participé à la rédaction de ces articles.

Chapitre II. Urriago J.D. Himmelman J.H., Gaymer CF., 2010. Responses ofthe black sea urchin Tetrapygus niger to its sea star predators Heliaster helianthus and Meyenaster gelatinosus under field conditions. JEMBE (soumis) Chapitre III. Urriago J.D. Himmelman J.H., Gaymer CF., 2010. Does the distribution of sea urchins Tetrapygus niger on elevated surfaces represent a strategy for avoiding prédation by sea stars? En préparation.

Les résultats obtenus au cours de cette thèse ont été présentés lors de congrès dont la liste apparaît ci-dessous: Urriago J.D., Himmelman J.H., Gaymer CF., 2010. Responses of the black sea urchin Tetrapygus niger to its sea star predators Heliaster helianthus and Meyenaster gelatinosus. Canadian Society of Zoology. Vancouver, Canada. Urriago J.D., Himmelman J.H., Gaymer CF., 2010. Responses of the black sea urchin Tetrapygus niger to its sea star predators Heliaster helianthus and Meyenaster gelatinosus. The 5th Annual meeting of the Canadian Society for Ecology and Evolution. Quebec City, Canada. Urriago J.D., Himmelman J.H., Gaymer CF., 2010. Too close for comfort: distance detection of predators by sea urchins. lere édition du Colloque du Département de biologie de l'Université Laval. Quebec City, Canada. Urriago J.D., Himmelman J.H., Gaymer CF., 2009. Too close for comfort: distance detection of predators by sea urchins. Forum québécois en sciences de la mer. Rimouski, Canada. IV

First of all I would like to thank my director John Himmelman who has been like a father to me in this country so far from home. He has been an excellent academic advisor, who during these 3 years in Canada has taught me with, patience and wisdom, a clear way to do research. My thanks for your advice, dedication, friendship and support. I sincerely thank my co-director Carlos Gaymer for his trust and support in my work. His patience and relevant inputs during the development of my master's research has been essential during the two field seasons (2008 - 2009) and during the previous year working on different Ecology and Marine Conservation projects (2007) in Chile. I especially want to thank him for his sincere friendship and the good energy extended to me since we met. I want to especially thank Clément Dumont, my friend and future Ph.D. co-director in Hong Kong, for giving me my first job as a marine biologist. He introduced me to Carlos and then suggested John as a potential Master's supervisor. Clement and Carlos are two more members ofthe extended Himmelman family which I have had the privilege to join as John's last student. This work would not have been possible without the important logistical support of my Elite diving team Etienne Renaud-Roy and Mayra Natalia Munoz Pinilla. Thanks for sharing with me continuous days of diving in the beautiful, cold, rough Chilean waters. I cannot leave out my grandfather Dario who also helped me logistically during several days of diving in Chile. The content of this thesis has been improved by the pertinent comments of my evaluation committee Ladd Johnson and Helga Guderley. Thanks to Ladd for his support and for shared pleasant moments in Quebec City and its surroundings. To Helga, the beautiful woman with an eternal smile and nice energy, thanks for her maternal friendship. I appreciated sharing Helga and John's love for nature and healthy living. Merci à Hélène Crépeau qui à été une personne clé dans la vérification des analyses statistiques. Je remercie Flavienne Bruyant pour les corrections de texte en Français. Je remercie le CRSNG (Canada) et le FONDECYT (Chile) pour leur soutien financier et logistique au projet. J'ai bénéficié pour ma part de bourses du Gouvernement du Québec, de Québec-Océan et du Département de Biologie de l'Université Laval. I have been surrounded by a wonderful and large group of friends (Biology Department) during my masters, especially my roommate Nicolas Martin ("mon petite c") who introduced me to his family in Shawinigan Sud and with whom in one way or another I shared thousands of experiences. I cannot put aside the many evenings of good coffee, accompanied by chess and great friends. Among them are: Hernân Pérez ("el especial"), the fancy English Samuel Collin ("my p. cow") a man of confidence, special musical preferences and with whom I shared really good times. Special thanks to David Pâez (liAustralianus maximus colombiensis") for his contributions and reviews on the articles in this thesis, but especially since my arrival to share good chess, good coffee, hard workouts in the gym and sauna evenings to discuss life. Finally, I specially want to thanks my mom Cecilia, my dad Dario and my grandmother Ruth for giving me unconditional support and trust. Thanks to them for encouraging me with tons of love which has kept me motivated. They will be forever in my heart. VI

Table des matières Page

CHAPITRE I. Introduction générale 1

CHAPITRE IL Responses of the black sea urchin Tetrapygus niger to its sea-star predators Heliaster helianthus and Meyenaster gelatinosus under field conditions 5

Résumé 6 Abstract 7 Introduction 8 Methods 11 Responses to different sea stars 11 Responses relative to the level of risk of predatory sea stars 12 Density effects on the sea urchin's response 12 Statistical analyses 13 Results 14 Responses to different sea stars 14 Responses relative to the level of risk of predatory sea stars 16 Density effects on the sea urchin's response 18 Discussion 21

CHAPITRE III. Does the distribution of sea urchins Tetrapygus niger on elevated surfaces represent a strategy for avoiding prédation by sea stars? 26

Résumé 27 Abstract 28 Introduction 29 Methods 31 Vil

Distribution of sea urchins and sea stars 31 Responses of sea urchins to sea stars on different types of bottom 32 Sustained (simulated) attacks on vertical walls 32 Sustained attacks on aggregations 33 Survival on high and low surfaces 33 Statistical analyses 34 Results 34 Distribution of sea urchins and sea stars 34 Responses of sea urchins to sea stars on different types of bottom 35 Sustained (simulated) attacks on vertical walls 35 Sustained attacks on aggregations 38 Survival on high and low surfaces 40 Discussion 43

CHAPITRE IV. Conclusion générale 48

Références bibliographiques 50 1

CHAPITRE I Introduction générale La prédation est une interaction entre espèces au cours de laquelle une des espèces, le prédateur, se nourrit de l'autre espèce, la proie. La proie n'est pas nécessairement tuée au cours de la prédation. Les prédateurs comprennent les herbivores, les carnivores et les parasites (Krebs, 2009). La prédation peut indirectement affecter le comportement de la proie, sa distribution, son abondance ainsi que la structure de sa communauté (Paine, 1980; Kerfoot et Sih, 1987; Estes et al., 1998; Broom et al., 2010). Les espèces proies présentent des réponses comportementales évoluées et diverses à l'égard des prédateurs. Ces réponses peuvent être groupées en deux catégories, l'évitement et la fuite (Lima et Dill, 1990). L'évitement minimise les rencontres avec les prédateurs, il s'agit par exemple des réponses comme le camouflage, l'aposématisme, la présence d'armures de protection ou bien des mécanismes de défense chimique (Nelson et Vance, 1979; Cronin et Hay, 1996; Saporito et al., 2006). La fuite tend à réduire la probabilité de blessures ou de mort après l'éventualité d'une rencontre avec un prédateur (Feder, 1963; Legault et Himmelman, 1993; Markowska et Kidawa, 2007; Gaymer et Himmelman, 2008). Dans le milieu marin la chimiodétection (à distance ou bien au contact) est un facteur important dans la médiation des interactions proie-prédateur. Les prédateurs peuvent souvent détecter l'odeur des espèces proies (McClintock et al., 1984; Rochette et al., 1994; Dale, 1997; Thompson et al., 2005), de même, les proies peuvent souvent détecter certaines substances excrétées par les prédateurs (Alexander et Covich, 1991; Covich et al., 1994; Svensen et Kiorboe, 2000; Markowska et Kidawa, 2007). Par exemple, les substances chimiques excrétées par les étoiles de mer prédatrices (asterisaponins; Gameau et al., 1989) déclenchent fréquemment des réponses défensives de la part de leur proies (Kats et Dill, 1998; McClintock et al., 2008b). Les espèces proies peuvent alors larguer dans le milieu naturel des substances chimiques qui alertent leurs congénères d'un danger potentiel (Smith, 1992; Chivers et Smith, 1998). Certains oursins de mer fuient lorsque exposés à des substances chimiques provenant d'étoiles de mer ou de congénères blessés (Dayton, 1975; Scheibling et Hamm, 1991). La plupart des études examinant la réponse des oursins aux risques de prédation ont été conduite en laboratoire, dans des conditions d'eau calme (Jensen, 1966; Bernstein et al., 1981; Tegner et Levin, 1983; Mann et al., 1984; Scheibling et Hamm, 1991; Legault et Himmelman, 1993; Hagen et Mann, 1994; Rodriguez et Ojeda, 1998; Matassa, 2010) et de courant unidirectionnel (Phillips, 1978; Moitoza et Phillips, 1979; Campbell et al., 2001; Hagen et al., 2002; Nishizaki et Ackerman, 2005). Les études réalisées sur le terrain l'ont généralement été dans des conditions de courant unidirectionnel ou en eaux calmes (Snyder et Snyder, 1970; Rosenthal et Chess, 1972; Dayton et al., 1977; Bernstein et al., 1983; Parker et Shulman, 1986; Vadas et al., 1986; Andrew et Macdiarmid, 1991; Vadas et Elner, 2003). Une seule étude rapporte les réponses des patelles au contact d'étoiles de mer prédatrices dans des conditions agitées (Espoz et Castilla, 2000). Aucune étude dans le milieu naturel n'examine la chimiodétection à distance dans des conditions agitées (présence de vagues), alors qu'en laboratoire, une étude par Gagnon et al. (2003) montre que l'étoile de mer Asterias vulgaris détecte sa proie, les moules, et se dirige vers elle dans un bac soumis à des vagues. La présente étude évalue les réponses de l'oursin de mer noir Tetrapygus niger envers deux de ces principaux prédateurs, les étoiles de mer Heliaster helianthus et Meyenaster gelatinosus. La présence de vagues est une caractéristique constante de la plupart des habitats où l'on trouve T. niger dans le centre et le Nord du Chili. Dans cette région, T. niger est aussi l'oursin •y le plus abondant (-40 ind. m" ; Vasquez et Buschmann, 1997). Les activités de broutage intensif de T. niger ont transformé de nombreuses zones sublittorales en zone dénudées. T. niger limite aussi souvent la distribution de la laminaire aux zones de faibles profondeurs et a provoqué des extinctions locales de la laminaire sublittoral Macrocystis integrifolia (Vega et al., 2005). Des études faites dans d'autres parties du monde ont également démontré que le broutage intensif par T. niger a provoqué le dénuement de zones marines et réduit la diversité et la biomasse des macroalgues (Himmelman et al., 1983; McClanahan et Shafir, 1990; Alcoverro et Mariani, 2002; Shears et Babcock, 2002). Les prédateurs peuvent affecter le comportement, la densité et la structure des populations des oursins de mer (Tegner et Levin, 1983; Sala et al., 1998; Tuya et al., 2004; Guidetti, 2006). La plupart des études sur les prédateurs des oursins se concentrent sur les étoiles de mer (Jensen, 1966; Rosenthal et Chess, 1972; Dayton et al., 1977; Moitoza et Phillips, 1979; Legault et Himmelman, 1993; Rodriguez et Ojeda, 1998; Hagen et al., 2002). Cependant, d'autre types de prédateurs, comme par exemple les loutres de mer (Estes et al., 1998), les poissons (Sala, 1997), les homards (Andrew et Macdiarmid, 1991) et les crabes (Scheibling et Hamm, 1991), ont été identifiés comme se nourrissant sur les populations d'oursins. De nombreux prédateurs sont identifiés comme consommateurs de T. niger, incluant les poissons Semicossyphus maculatus (Fuentes, 1981), Graus nigra (Fuentes, 1982), (Rodriguez et Ojeda, 1998), Cheilodactylus variegatus et Oplegnathus insignis (Medina et al., 2004), et l'étoile de mer Luidia magellanica (Gaymer et Himmelman, 2008), cependant, les deux prédateurs les plus fréquemment observés se nourrissant de T. niger sont les étoiles de mer H. helianthus et M. gelatinosus (Barrios et al., 2008; Gaymer et Himmelman, 2008). Ces étoiles de mer ont été décrites comme des espèces prédateurs clés pour les communautés rocheuses peu profondes du Nord du Chili (Gaymer et Himmelman, 2008; Barahona et Navarrete, 2010). H. helianthus est un prédateur ubiquiste qui consomme ses proies en fonction de leur disponibilité (Gaymer et Himmelman, 2008; Barahona et Navarrete, 2010), alors que M. gelatinosus est un prédateur sélectif qui préfère consommer T. niger (Gaymer et Himmelman, 2008). Dans le Nord du Chili la densité moyenne de H. helianthus est de 3 ind. m" et celle de M. gelatinosus est <0.5

■y ind. m" . Les deux espèces sont endémiques le long de la côte Ouest de l'Amérique du Sud (Dayton et al., 1977; Tokeshi et al., 1989; Gaymer et Himmelman, 2008; Navarrete et Manzur, 2008). H. helianthus possède un corps robuste et aplati avec de nombreux bras (jusqu'à 40), elle peut atteindre 32 cm de diamètre (Tokeshi et al., 1989). En revanche, M. gelatinosus a un corps mou avec 6 larges bras et peut atteindre jusqu'à 56 cm de diamètre (Dayton et al., 1977; Gaymer et Himmelman, 2008). Deux études se concentrent sur les réponses comportementales de T. niger envers ses étoiles de mer prédatrices. Dayton et al. (1977) ont examiné les réponses de plusieurs invertébrés à la présence de M. gelatinosus dans des bassins créés par la marée; ils se sont particulièrement concentrés sur les réponses de la variété commerciale d'oursin . Ils mentionnent que T. niger est capable de détecter M. gelatinosus à une distance de 50 à 130 cm et de répondre presque immédiatement par une fuite vers des surfaces élevées. Rodriguez et Ojeda (1998) ont réalisé des études dans des bacs de laboratoire pour examiner la réponse des oursins à la présence de M. gelatinosus et à celle du poisson P. chilensis. Ils rapportent que T. niger augmente son taux de déplacement lorsque ces prédateurs sont proches et qu'une proportion plus grande d'individus répond lors des tests avec M. gelatinosus (92 %) que lors des tests avec P. chilensis (67 %). L'objectif général de cette thèse est d'examiner les réponses comportementales de l'oursin de mer noir Tetrapygus niger face aux étoiles de mer Heliaster helianthus et Meyenaster gelatinosus, dans les conditions naturelles. J'ai conduit des expériences dans le milieu naturel sous des conditions agitées (vagues) conditions prédominantes dans l'habitat de ces espèces dans le Nord du Chili. Dans le chapitre II, j'ai examiné les réponses de T. niger à des degrés variés de proximité avec ses deux prédateurs. Plus spécifiquement j'ai (1) évalué la capacité de l'oursin à différencier entre deux étoiles de mer prédatrices et une non prédatrice, (2) enregistré les réponses comportementales de l'oursin soumis à des attaques simulées de H. helianthus et M. gelatinosus, (3) déterminé la distance de réaction de l'oursin face à ses deux étoiles de mer prédatrices et (4) étudié dans quelle mesure la densité de congénères modifie les réponses des oursins envers leurs prédateurs, ce qui suggérerait l'utilisation de signaux d'alarme. Dans le chapitre III, j'ai regardé si la micro-distribution de T. niger sur les surfaces surélevées réduisait les risques d'attaque par les étoiles de mer prédatrices H. helianthus et M. gelatinosus. Spécifiquement j'ai (1) examiné l'association des oursins et des étoiles de mer sur les surfaces en hauteur et plus basses, (2) examiné les réponses des oursins aux étoiles de mer dans différentes situations et (3) réalisé des essais avec des oursins entravés pour comparer les risques de prédations sur les surfaces surélevées versus les surfaces plus basses. CHAPITRE II

Responses of the black sea urchin Tetrapygus niger to its sea-star predators Heliaster helianthus and Meyenaster gelatinosus under field conditions RESUME Nous avons utilisé des expériences dans le milieu naturel pour étudier les réponses de l'oursin de mer noir Tetrapygus niger à la prédation des étoiles de mer. Des essais impliquant des attaques simulées (un ou plusieurs bras d'étoile de mer placés sur une moitié de l'oursin) ont montrés que les oursins étaient capables de différencier les étoiles de mer prédatrices Heliaster helianthus and Meyenaster gelatinosus des étoiles de mer non-prédatrices , et qu'ils ne montraient presque aucune réponse dans le cas d'une fausse étoile de mer. Les réponses des oursins à différents niveaux de menace représentés par les deux étoiles de mer prédatrices ont également été comparées. Le plus haut niveau de menace est l'attaque simulée, puis un simple contact, et enfin des étoiles de mer placées à différentes distances de l'oursin. Tous les oursins ont réagis aux attaques simulées et à la mise en contact avec les deux espèces d'étoiles de mer. La proportion de réponse diminue avec la distance et ce plus rapidement pour H. helianthus (0 % à une distance de 30 cm) que pour M. gelatinosus (33 % à une distance de 50 cm). A chaque niveau de menace présentant une réponse pour chacune des espèces prédatrices, les oursins ont répondus plus rapidement dans le cas de M. gelatinosus que dans celui de H. helianthus. Dans une troisième expérience, une étoile de mer prédatrice était ajoutée dans une aire circulaire (1-m de diamètre) dans laquelle 4 à 8 ou 11 à 19 oursins (non dérangés) étaient présents. Les oursins fuyaient la zone plus rapidement dans le cas de M. gelatinosus, mais le taux de fuite ne variait pas avec la densité (comme ce serait le cas s'il y avait communication entre les oursins par des signaux d'alarme). Nos observations suggèrent que M. gelatinosus représente une menace de prédation plus importante que H. helianthus. Ceci est en accord avec des observations faites dans le milieu naturel montrant que les oursins sont consommés plus fréquemment par M. gelatinosus. Ces expériences en milieu naturel sont les premières démontrant la chimiodétection à distance chez un invertébré marin sous des conditions de courant bidirectionnel induit par les vagues. ABSTRACT We ran field experiments to examine the responses of the black sea urchin Tetrapygus niger to predatory sea stars. Trials involving simulated attacks (one or several arms of a sea star being placed on top of half the urchin) showed that the urchin differentiated between the predatory sea stars, Heliaster helianthus and Meyenaster gelatinosus, and a non-predatory sea star, Stichaster striatus, and showed almost no response to a sea star mimic. We further compared the responses ofthe urchin to different threat levels presented by the two predatory sea stars. The highest threat level was a simulated attack, then mere contact, and subsequently sea stars being placed at different distances from the urchin. All urchins responded to simulated attacks and contact with both sea stars. The proportion responding decreased with distance and more rapidly in trials with H. helianthus (0 % at a distance of 30 cm) than with M. gelatinosus (33 % at a distance of 50 cm). At each of the threat levels where there was a response to both sea stars, the urchins responded more rapidly to M. gelatinosus than to H. helianthus. In a third experiment where a predatory sea star was added to a circular area (1-m diameter) in which either 4-8 or 11- 19 undisturbed urchins were present, the urchins fled the area more rapidly when the added sea star was M. gelatinosus, but the rate of fleeing did not vary with density, as might occur if there was communication among urchins using alarm signals. Our observations suggest that M. gelatinosus presents a stronger predatory threat than H. helianthus. This corresponds to field observations showing that the urchins are more frequently consumed by M. gelatinosus. These are the first field experiments demonstrating distance chemodetection by a marine invertebrate under back-and-forth water flow from wave activity. 8

INTRODUCTION Prédation is an important ecological factor because of its effects on prey behaviour and survival, and ultimately community structure (Paine, 1980; Estes et al., 1998; Gaymer and Himmelman, 2008). Natural selection should lead predators to select prey that increase their fitness and lead prey to develop defenses that decrease the risk of being eaten (Kerfoot and Sih, 1987). The diverse behavioural responses to predators (Legault and Himmelman, 1993; Sih and Wooster, 1994; Rochette et al., 1998) can be grouped into two categories, avoidance and escape adaptations (Lima and Dill, 1990). Avoidance adaptations act to reduce encounters with predators, for example cryptic and aposematic coloration, protective armor and chemical defenses (Cronin and Hay, 1996; Terlau et al., 1996; Gaymer et al., 2001a; 2001b). Escape adaptations minimize the likelihood of death when a predator is encountered, for example flight responses upon contact with predators and morphologies that reduce the predator's handling efficiency (Jensen, 1966; Vadas, 1977; Bernstein et al., 1983; Miller and Byrne, 2000; Furrow et al., 2003; McClintock et al., 2008b). Chemodetection is often involved in interactions among marine . For example, numerous studies show that sea stars can detect substances exuded by potential prey (McClintock et al., 1984; Rochette et al., 1994; Dale, 1997) and that prey can detect the odours from predators (Phillips, 1978; Vadas et al., 1986; Alexander and Covich, 1991; Covich et al., 1994; Svensen and Kiorboe, 2000; Markowska and Kidawa, 2007). Chemical signals from predatory sea stars often trigger particular defensive responses (Kats and Dill, 1998; Nishizaki and Ackerman, 2005) and in some systems chemicals released by prey being attacked alert conspecifics to the danger (Smith, 1992; Chivers and Smith, 1998). For example, numerous asteroids, ophiuroids and echinoids exhibit defensive behaviours when exposed to chemicals from predators or injured conspecifics (Snyder and Snyder, 1970; Dayton, 1975; Parker and Shulman, 1986; Vadas et al., 1986; Scheibling and Hamm, 1991; Legault and Himmelman, 1993; Rodriguez and Ojeda, 1998; Rosenberg and Selander, 2000). A few studies also demonstrate alarm signaling among prey species in response to an attack by a common predator species (Brown and Godin, 1997). All species within a prey guild should benefit from such alarm signals (thus have a reduced risk of prédation), independent of the species that produces the signal (Crawl and Covich, 1990; Alexander and Covich, 1991; Covich et al., 1994; McClintock et al., 2008a). A number of studies report that escape responses by sea urchins can reduce prédation risk (Snyder and Snyder, 1970; Bernstein et al., 1981; Bernstein et al., 1983; Parker and Shulman, 1986; Vadas et al., 1986; Scheibling and Hamm, 1991; Hagen and Mann, 1994). Several studies suggest that sea urchins form aggregations that decrease the vulnerability to predators (Garnick, 1978; Mann, 1982; Bernstein et al., 1983). However, this hypothesis is contested because the aggregative behaviour may be related to the patchy distribution of food (Vadas et al., 1986; Himmelman and Nedelec, 1990; Scheibling and Hamm, 1991; Hagen and Mann, 1994; Rodriguez and Ojeda, 1998; Vadas and Elner, 2003), bottom topography (Laur et al., 1986), or juveniles seeking protection under adults (Nishizaki and Ackerman, 2005). Most experiments examining contact or distance chemodetection in predator-prey interactions among marine organisms have been ran in the laboratory (Lima and Dill, 1990; Chivers and Smith, 1998; Drolet and Himmelman, 2004; Jackson and Kiorboe, 2004; Thompson et al., 2005; McClintock et al., 2008a; Himmelman et al., 2009). Also, the studies examining responses of sea urchins to predatory risk have been mainly conducted under still water conditions (Jensen, 1966; Bernstein et al., 1981; Tegner and Levin, 1983; Mann et al., 1984; Scheibling and Hamm, 1991; Legault and Himmelman, 1993; Hagen and Mann, 1994; Rodriguez and Ojeda, 1998; Matassa, 2010) or unidirectional flow (Phillips, 1978; Moitoza and Phillips, 1979; Campbell et al., 2001; Hagen et al., 2002; Nishizaki and Ackerman, 2005). Most of the studies made in the field were executed under unidirectional flow or calm conditions (Snyder and Snyder, 1970; Rosenthal and Chess, 1972; Dayton et al., 1977; Bernstein et al., 1983; Parker and Shulman, 1986; Vadas et al., 1986; Andrew and Macdiarmid, 1991; Vadas and Elner, 2003). An ability to identify the risk associated with different predatory sea stars has been shown for a variety of animals, including gastropods (Feder, 1963, 1972; Phillips, 1977; Fishlyn and Phillips, 1980; Rochette et al., 1998; Mahon et al., 2002; Markowska and Kidawa, 2007), cnidarians (Weightman and Arsenault, 2002) and sea stars that are prey (McClintock et al., 2008b). These species usually flee when there is contact with the predatory sea star or detection from a distance. One study executed in northern Chile reports escape responses of intertidal limpets resulting from contact with sea stars under wave conditions (Espoz and Castilla, 2000). No studies have examined distance chemodetection under wave conditions in the field, although one laboratory study (Gagnon et al., 2003) shows that the sea star Asterias vulgaris detects and moves towards food (mussels) in a wave tank. 10

Wave surge is a consistent feature of most habitats of the sea urchin Tetrapygus niger in central and northern Chile. This sea urchin is the most conspicuous benthic grazer in Chile (ranging from 2 to 85 ind. m"2; Rodriguez, 2003) and its intensive grazing causes the formation of extensive urchin barrens (Vasquez and Buschmann, 1997). The two predators most frequently observed feeding on T. niger are the sea stars Meyenaster gelatinosus and Heliaster helianthus (Barrios et al., 2008; Gaymer and Himmelman, 2008). These sea stars have been described as keystone predators in shallow rocky communities in northern Chile (Gaymer and Himmelman, 2008; Barahona and Navarrete, 2010). The multi-armed sea star H. helianthus is a generalist predator that consumes prey according to availability (Gaymer and Himmelman, 2008; Barahona and Navarrete, 2010), whereas M. gelatinosus is a highly selective predator that prefers the urchin T niger and the sea star H. helianthus (Gaymer and Himmelman, 2008). Dayton et al. (1977) examined responses of a variety of invertebrates to M. gelatinosus in tide pools, but focused on the responses ofthe commercially-harvested urchin Loxechinus albus. They mention that T. niger could detect M. gelatinosus at a distance (50 to 130 cm) and responded within seconds with a fleeing response. Laboratory studies by Rodriguez and Ojeda (1998) showed that the addition of M. gelatinosus or the predatory fish Pinguipes chilensis to tanks caused T. niger to move faster, but not to aggregate. The proportion of urchins showing increased movement was higher in the trials with M. gelatinosus (92 %) than in those with Pinguipes chilensis (67 %) (Rodriguez and Ojeda, 1998). The present study examines the responses of the black sea urchin T niger to varying degrees of proximity with the predatory sea stars M. gelatinosus and H helianthus, representing different levels of predatory risk. All trials were made under field conditions involving continuous back-and-forth wave movement. Specifically, we (1) examined the ability of the urchin to discriminate different sea stars, two predators and a non predator, (2) evaluated the behavioural responses of the urchin to simulated attacks by, and to contact with, the two predatory sea stars, (3) determined the distance over which the urchin reacts to the two predatory sea stars and (4) investigated whether the density of conspecifics modifies the responses of the urchin to its predators, which would suggest the use of alarm responses. 11

METHODS Our study was conducted during June through August in 2008 and 2009 in the shallow subtidal zones at Cisnes (27°14'50"S, 70°57'34"W) and La Herradura (29°28'r'S, 71°21'18"W) Bays in northern Chile. All manipulations were made using SCUBA diving in wave-exposed environments between 2 and 10 m in depth. The two sites were moderately sloped bottoms (down to -10 m depth and at -30 m from shore) that supported rocky barrens communities dominated by the sea urchin Tetrapygus niger. At both sites the sea stars Heliaster helianthus and Meyenaster gelatinosus were present at densities of <3 m2. During our trials, water temperature ranged from 12 to 14 °C In all experiments, divers collected sea stars randomly from among individuals that were stationary and not feeding. In each trial different sea stars, urchins and urchins aggregations were used.

Responses to different sea stars Our first experiment was ran on a relatively smooth horizontal platform (bedrock with few surface irregularities) at La Herradura Bay in 2008 and examined the ability of the urchin to differentiate among different sea stars representing varying degrees of predatory risk. For this, we recorded the responses of isolated urchins (5-6 cm in diameter) to simulated attacks by H. helianthus (9-14 cm in radius) and M. gelatinosus (13-21 cm), as well as by the sea star Stichaster striatus (8-11 cm), which does not prey on urchins (Viviani, 1978), and by sea star mimics (5 balloons filled with sand and covered with synthetic leather, -10 cm in radius). For each simulated attack, the sea star (one arm in attacks with M. gelatinosus, S. striatus and the mimic, and several arms in attacks with H. helianthus) was placed on top of half the target urchin (the target urchin was separated from other urchins at least 20 cm) and held in this position for up to 120 s. The sea star was maintained in the initial position, even as the urchin moved away. Preliminary trials indicated that if there were a response it would occur within 120 s. For each attack, we recorded if and when the urchin (1) raised its spines, (2) started displacing, and (3) severed contact with the sea star. In this way, 20-26 simulated attacks were made for each of the three sea stars and the mimic. The proportion of urchins responding and reaction times for each ofthe three parameters were analyzed separately. The same body-size range oiH helianthus, M. gelatinosus and T. niger was used in subsequent experiments. 12

Responses relative to the level of risk of predatory sea stars We evaluated the response time of undisturbed urchins (time until the urchin displaced itself) subjected to situations representing different levels of predatory risk by H. helianthus and M. gelatinosus. The highest level of risk was to simulated attacks (as described above) and the next level was to mere contact with the tip of one arm of the sea star (we touched spines on the sides of the urchin). The subsequent risk levels involved placing sea stars at different distances from the urchin. The first distance was 10 cm and then distances were increased by 10-cm intervals until the urchins no longer responded, or up to a maximum of 50 cm. In each trial the sea star was placed at the desired distance along the axis of the wave activity and held there until the urchin responded or for a maximum of 120 s. For each sea star we made 40 trials with simulated attacks, 14-15 trials with mere contact, and 14-15 times for each ofthe distances (new sea stars and urchins were used in every trial at each threat level, including at each distance). These experiments were conducted on a relatively smooth horizontal bedrock platform at Cisnes Bay in August 2009.

Density effects on the sea urchin's response We conducted experiments in 2008 at Cisnes Bay aimed at determining if the density of conspecifics affects the urchin's behavioural responses to the two predatory sea stars. We recorded the time to respond to a sea star being placed nearby for urchins at low or high densities. In each trial, we first placed a 1-m diameter circular hoop (made with 5-mm thick wire) around urchins found on a relatively flat rocky platform (we chose areas where there were no urchins in the center of the hoop) and removed urchins when necessary so that there were either 4-8 (low density) or 11-19 (high density) urchins within the area defined by the hoop. Then we placed a sea star, collected haphazardly from the surrounding area, in the empty space at the center of the hoop (the sea star was always at least 5 cm from the urchins) and recorded the departure times for urchins leaving the area defined by the hoop during a 6-min period. The urchins readily climbed over the wire hoop. The sea star was maintained at the center of the hoop during each trial. This procedure was repeated 5 times with M. gelatinosus and H. helianthus, and 8 times without sea stars (control trials), at both low and high densities of T. niger (each trial was on a new experimental area). An increased response at a higher density would imply that the urchins were responding to some sort of alarm signal. 13

Statistical analyses For the experiments evaluating the ability of the urchin to differentiate between different sea stars representing varying degrees of predatory risk, we quantified the proportion of urchins showing a response to the sea stars and, response times, for the three variables, (1) raising the spines, (2) active displacement and (3) severing contact. A £2 test was used to test if the proportion of urchins responding to the different sea stars and the mimic varied. When the global test was significant we followed with pairwise comparisons using the same test. A one-way ANOVA was performed to compare the mean time of urchins responding to the different sea stars and the mimic and followed with pairwise comparisons using protected Fisher least square difference tests (LSD). These two analyses were performed separately for each variable (spines raising, displacement and severing contact). Data were log transformed when necessary to meet the assumptions of normality and homogeneity of variance. Normality was tested using the Shapiro-Wilk's test (SAS, 2008) and homogeneity of variances using the Levene test (Snedecor and Cochran, 1989). A generalized linear model (using the GENMOD procedure with the binomial distribution and the logit link was used in SAS, 2008) was then used to compare the proportion of urchins responding to different levels of predatory risk represented by the two predatory sea stars. In this model there were two fixed factors, the sea star and the level of predatory risk. This analysis was only applied to the data for distances of 20 and 30 cm because the urchin responses to the two sea stars were the same (100%) in simulated attacks, with mere contact and at 10 cm and the trials for distances of 40 and 50 cm we performed only with M. gelatinosus. A two-way ANOVA, with the same fixed factors, was used to compare the urchin response time to situations representing different levels of predatory risk. The MIXED procedure (SAS, 2008) was used to model the heterogeneity observed in each combination of the factors. In this case, no transformation was necessary to meet the normality assumption. To test if the mean response time of urchins varied in the trials with different levels of predatory risk from M. gelatinosus, a one-way ANOVA was performed. The fixed factor was predatory risk. The MIXED procedure was also used because of the heterogeneity observed in each combination of predatory risk test (SAS, 2008). The log- transformation was used to meet the normality assumption. To determine if the density of conspecifics modified the responses of the urchin to M. gelatinosus and H. helianthus, we first compared the proportion of urchins leaving the circular 14 area using a generalized linear model (GENMOD procedure as described above). The two fixed factors were the sea star species and urchin density. Then, a conditional analysis was applied to the time when the urchins left the area defined by the hoop. Trials in which no urchins left the hoop were not included in the analysis. A two-way ANOVA was used to compare the effect of urchin density and the sea star species. No transformation was necessary.

RESULTS Responses to different sea stars The simulated attacks with three species of sea stars and the mimic sea star showed that the sea urchin Tetrapygus niger differentiates between predatory and non predatory sea stars (Fig. 2.1). The urchin's first response to Heliaster helianthus and Meyenaster gelatinosus was raising the spines, which occurred after about 3 s of exposure. Displacement followed after about 7 s and the displacement resulted in severing of contact after about 28 s. The urchins used their spines and podia to move away from the sea stars. There were no differences in the numbers of urchins responding and response times for the trials with the two sea stars (Fig. 2.1; /*>0.05). The time until contact was severed differed significantly (j°=0.038), although the mean time differed by only 21%. In contrast, few urchins responded to the sea star mimic and reaction times were much greater, as only six urchins (30 %) showed displacement within the 120-s observation period and only one of these (5 %) raised its spines and severed contact with the mimic. Responses to the non-predatory sea star S. striatus were intermediate. Although 88 % of the urchins raised their spines and displaced themselves, and 48 % severed contact with S. striatus, the reaction times were almost as slow as in the trials with the mimics (Fig. 2.1). In 2009, we also made trials to record the time required for the urchin to extend its podia. The responses did not vary between the two predatory sea stars, either for the number of urchins responding (all responded) or the response time (P=0.84). Podia extension took place in about 8.7 s in response to attacks by H. helianthus and in 8.5 s in attacks by M. gelatinosus, thus at almost the same time as when displacement began in the trials with simulated attacks made in 2008. 15

100

"5 80 - c o a 60 a c o 40 c B o a 8 20 - a.

0 353 %&

n=26 n H. helianthus 100 n = 26 m M. gelatinosus " = 25 os.striatus 80 n = 20 EiMimic '■ i 60 o l 40 * m o mm i a. o 20 m cr I I i Raising spines Displacement Severing contact

Figure 2.1. Proportion (%) of sea urchins (Tetrapygus niger) responding within 120 s and response time for three variables (spines raising, displacement and severing contact) used to evaluate the urchin's response to simulated attacks by the sea stars Heliaster helianthus, Meyenaster gelatinosus, Stichaster striatus and a mimic sea star. Values are means ± SE. Each variable was analyzed separately; bars not sharing the same letter are different (P<0.05). 16

Responses relative to the level of risk of predatory sea stars All urchins moved away in the trials representing to the first three levels of risk (simulated attacks, mere contact and a distance of 10 cm) ofthe two predatory sea stars (Fig. 2.2). Then, with further increases in distance (and decreased risk) the percent displacing declined. The decrease was more rapid in the trials with H. helianthus than with M. gelatinosus, for example, no urchins responded when H. helianthus was placed at 30 cm from the urchin, whereas 33.3 % ofthe urchins still responded in the trials in which M. gelatinosus was placed at 50 cm. To further define the range of detection of H. helianthus, we performed additional trials at 25 cm and found that 29 % ofthe urchins (6/21) detected the sea star (mean response time 43.3 s, SE±13.0). In the simulated attacks, there was no difference in response time in the trials with the two predatory sea stars (P=0.16). In contrast, with mere contact the response time was notably more delayed in the trials with H. helianthus than with M. gelatinosus (P<0.01). At distances of 10 and 20 cm response times to the two sea stars were similar (P>0.05). The response time in trials with M. gelatinosus was notably higher at 50 cm than at closer distances, but the increase was not significant (P>0.05). 17

Distance chemodetection

— 100 >^^^^ JS • • S. M. gelatinosus g 80 «e^ a | 60 c o a. S 40 H. helianthus c o t â 20 S a n i 1— 1 1 —■ 1 1 p=0.3 pO.0001 80 H. helianthus i M. gelatinosus ^ 60

«u p=0.06 40 p=0 30 8 pO.01 ao «/> î 20 p=0.16

c B

(40) (40) (15) (14) (15) I(15) I(15 ) (15) (15I) (15) (15) Simulated Mere 10 cm 20 cm 30 cm 40 cm 50 cm attack contact Decreasing threat level

Figure 2.2. Proportion (%) of sea urchins (Tetrapygus niger) responding within 120 s and response time in trials in which the urchin was exposed to the sea stars Heliaster helianthus and Meyenaster gelatinosus at different threat levels: simulated attacks, mere contact (with the tip of one arm ofthe sea star) and distances (10-50 cm) from the urchin. Values are means ± SE. The P values give the probability that the urchin responded in the same way to the two sea stars. For each sea star, bars not sharing the same letters are different (PO.05). The number of urchins used in each treatment is indicated in parenthesis. 18

Density effects on the sea urchin's response The final experiment in which we measured the fleeing of urchins from circular areas where we added a predatory sea star (H. helianthus or M. gelatinosus) or nothing (control), showed no effect of density but an effect ofthe treatments (Table 1). In fact, in the trials with each sea star, the percentage of urchins leaving the hoop and departure times were similar for the two densities of urchins (Table 1; Fig. 2.3). The rate of fleeing was much greater in the trials with M. gelatinosus than with H. helianthus. For example, >95 % of the urchins left the hoop within 4 min in the trials with M. gelatinosus, compared to <40 % at 6 min in the trials with H. helianthus. In the trials with H. helianthus, the urchin often only moved a short distance and not enough to leave the hoop; and very few urchins left the hoop near the end of the trials (between 4 and 6 min). In contrast, in the trials with M. gelatinosus almost all urchins (99%) moved outside the hoop within 6 min. In the control trials (where nothing was added) most urchins remained stationary or only moved slightly. Only -10 % ofthe urchins left the hoop in 6 min and some urchins from outside moved into the area defined by the hoop (Fig. 2.3). 19

100

Time (min)

Figure 2.3. Cumulative proportion of sea urchins (Tetrapygus niger) leaving a circular area at different time intervals after a sea star, Heliaster helianthus or Meyenaster gelatinosus, was added to the center area. The initial number of urchins in the area was 4-8 (low density) or 11-19 (high density). Values are means ± SE. At the end of the trials (at 6 min), the three treatments differed (PO.05). 20

Table 1. Logistic regressions for the proportion of sea urchins (Tetrapygus niger) leaving the circular areas defined by a 1 -m diameter hoop (GENMOD procedure; SAS, 2008). The variables were density (low and high) and the three treatments (the sea stars, Heliaster helianthus and Meyenaster gelatinosus and the control, without a sea star). Statistical differences in proportions were identified using Pearson x2 tests. Means were compared using LSD tests.

Source df x2 P-value Treatment 2 212.29 O.0001 Density 1 0.66 0.42 Treatment x Density 2 0.79 0.67 Control versus H helianthus 1 12.73 0.0004 Control versus M. gelatinosus 1 263.55 O.0001 H helianthus versus M. gelatinosus 1 240.87 O.0001 21

DISCUSSION Our study is the first to examine responses of the black sea urchin Tetrapygus niger to predators under field conditions. Given that the back-and-forth flow from waves is a prevalent feature of most habitats where T. niger is found along the coast of northern Chile, wave action should be considered in evaluating the urchin's responses to predators. Our field data demonstrate that T. niger can (1) differentiate between predatory and non-predatory sea stars, (2) distinguish between different threat levels presented by predatory sea stars, and (3) detect predatory sea stars at a distance. The behavioural responses of T. niger to simulated attacks by predatory sea stars consisted of raising the spines (3 s), followed by displacement and elongation of podia (both at about 8 s). The displacement led to severing contact with the predator (in about 30 s). We have also observed this sequence when sea stars naturally come into contact with sea urchins in the field, though we did not quantify reaction times. This sequence and similar reaction times have been reported for the sea urchin Strongylocentrotus droebachiensis in response to contact with its predators, the sea stars Leptasterias polaris and Crossaster papposus (Legault and Himmelman, 1993). The first response for that urchin was bending of spines away from the area of contact (occurring almost immediately as for T. niger), then elongation of podia and displacement (about 10 s later). They do not report the time when contact with the sea star was severed. In the field T. niger is likely to escape predatory sea stars unless its movement is blocked by other urchins or by bottom irregularities (Dayton et al., 1977). T. niger can more quickly sever contact with H. helianthus and M. gelatinosus on smooth horizontal platform than on irregular horizontal bottom (Chapter II). Any inability to flee generally allows the sea star to wrap its arms around the urchin and begin digesting it externally (JDU, observations). Our trials using simulated attacks demonstrate that T. niger differentiates between predatory and non-predatory sea stars. The response was strongest and similar for the two predatory sea stars. We recorded a 100 % response for the three variables evaluated (spine raising, displacement and severing of contact) in the trials with both H. helianthus and M. gelatinosus. Further, the reaction time of the urchins was rapid and similar for the two sea star predators, but the urchins tended to sever contact more rapidly with H. helianthus than with M. gelatinosus. Only a few urchins responded to the mimic sea star, and the reaction time for those that did respond was very slow. Most of the urchins remained in contact with the mimic. In the 22 trials with the non-predatory sea star, S. striatus, although a large proportion of urchins (88 %) raised their spines and displacement, only half (48 %) severed contact. Also, the response time to respond for all three variables was long, and indeed similar to those with a mimic. The ability of animals to differentiate between predators and non-predatory species has been demonstrated in many species (Sih et al., 1985; Chivers and Smith, 1998), including several species of urchins (Snyder and Snyder, 1970; Parker and Shulman, 1986; Hagen et al., 2002). The high proportion of urchins raising their spines and displacing in response to attacks by S. striatus is possibly because S. striatus, being an asteroid , shares morphological and some chemical characteristics with H helianthus and M. gelatinosus. However, the low proportion severing contact suggests that S. striatus represents a low risk of prédation. Other studies also report no response of urchins to mimic sea stars (Legault and Himmelman, 1993; Gaymer et al., 2002). In northern Chile, the responses of the limpet Lottia orbignyi to its predators parallel those of the urchins. It vigorously reacts to the sea star H helianthus, a known limpet consumer, but shows almost no reaction to S. striatus, which has never been reported to prey on limpets (Espoz and Castilla, 2000). M. gelatinosus is also a predator of H. helianthus, but field trials involving simulated attacks showed that H. helianthus takes 134 s to sever contact with M. gelatinosus (Gaymer and Himmelman, 2008). This was four times longer than the average time T. niger took to sever contact with M. gelatinosus (31 s). The slower response may be because H. helianthus can avoid total prédation by autotomizing arms, and thus can take higher risks in interactions with M. gelatinosus, or alternatively because H. helianthus is a slower moving . Further understanding of the responses of T. niger to its predators was provided by our field trials involving situations with difference levels of prédation risk. All urchins tested responded to the higher risk levels (attack and contact). Although the reaction time to simulated attacks was similar for the two sea stars, the reaction time to contact was much shorter in the trials with M. gelatinosus than those with H. helianthus. The high reaction time to simulated attacks by these two sea stars suggests that the urchin perceives such a high risk, irrespective of the predator, that it responds with the maximal response. The maintenance of a short reaction time in the trials with mere contact to M. gelatinosus suggests this species is consistently perceived as a high risk. In contrast, the slower response to mere contact with H. helianthus, compared to the response time in the simulated attacks, suggests that just touching this predator is perceived as a lower risk. The capacity of T. niger to distinguish between H. helianthus and M. 23 gelatinosus could decrease the cost of disrupted foraging activity and metabolic alterations resulting from defense and escape responses. The experiments where predators were placed at different distances provided further evidence of a stronger response to M. gelatinosus than to H. helianthus. In the trials with H. helianthus the percentage of urchins responding began to drop at a distance of 20 cm and was null at 30 cm (additional trials made at 25 cm, not shown in Fig. 2.2, show a 29 % response at this distance). In contrast, the major drop in the percentage of urchins responding to M. gelatinosus began at 40 cm, and 33 % of the urchins still responded at 50 cm, the furthest distance studied. Trials were not conducted at greater distances with M. gelatinosus because it was rare to find urchins separated from other urchins by distances of >50 cm. The different perception of the urchin to sea-star threats is also indicated by observations of the feeding of the two sea stars in the field. M. gelatinosus is a highly selective predator that strongly chooses echinoderm prey, such as T. niger and H. helianthus, whereas H. helianthus is a generalist feeder that consumes a wide variety of prey according to availability (Gaymer and Himmelman, 2008; Barahona and Navarrete, 2010). As a means of limiting energy expenditure, H. helianthus probably feeds on animals that are easier to attack than T. niger, as for example, the mussel Semimytilus algosus or the abundant gastropod Turritella cingulata (Gaymer and Himmelman, 2008). The weaker responses of T. niger to H. helianthus compared to M. gelatinosus could be one reason why urchins are often found in close proximity to H. helianthus but not to M. gelatinosus (JDU, observations). A closer association with H. helianthus than with M. gelatinosus was also indicated by the analysis of survey data involving sampling of 1-m2 quadrats at five locations (Gaymer, unpubl. data). There was a significant positive correlation between the density ofthe urchin and that oiH. helianthus but not with M. gelatinosus. Numerous previous studies have documented the ability of urchins, and other prey species, to detect predators at a distance, and some studies using extracts from predatory sea stars suggest urchins have the capacity to detect these chemicals (e.g., saponins; Phillips, 1978; Garneau et al., 1989). Our observations are useful because we evaluated the urchin's responses to intact sea stars in the field. In natural environments, wave action and associated turbulence would be expected to reduce the effectiveness of distance chemodetection (Jackson and Kiorboe, 2004). The ability of T. niger to detect and distinguish between predators, even at a distance and under wave surge, should permit it to decrease the probability of encounters. This may lead to changes 24 in its distribution and association with other species (e.g., T. niger appears to aggregate on elevated surfaces to escape from sea stars; Chapter III). We showed that urchins in groups disperse when H. helianthus or M. gelatinosus is placed nearby. The movement away from the predator was considerably more rapid in the trials with M. gelatinosus than in those with H. helianthus. However, the rate of dispersion did not vary with urchin density. During these experiments, we noted that the urchins closest to the sea star reacted faster than urchins at greater distances, although this was not quantified. Probably odours from the sea star were strongest for the closest urchins. Our field experiments support the laboratory observations by Rodriguez and Ojeda (1998) indicating that T. niger is able to recognize the presence of M. gelatinosus at a distance. They observed that urchins in 7-individual aggregations (large individuals as in our trials) reacted by moving away from the sea star; 19 % ofthe urchins tested remained in the vicinity 5 min after the introduction of M gelatinosus in the middle of a 90 x 100 cm tank (with a continuous inflow of sea water). In contrast, in our field study <3 % ofthe urchins in the 1-m diameter hoops remained in the hoop 5 min after the addition of M gelatinosus. Thus, although the experimental area, and the density and size of urchins were similar in our study and that by Rodriguez and Ojeda (1998), the urchins fled more rapidly in our study. Presumably the slower flight by the urchins in their study was due to the artificial conditions in the laboratory. The grazing by the high densities of the sea urchin T. niger in Chile has strong impacts on benthic communities and can cause extinction of beds in localized areas (Vega et al., 2005). Since kelp beds are important larval recruitment areas and nursery grounds for a number of commercial invertebrates and vertebrates, understanding the factors that regulate urchin populations is crucial. Several factors favor the high numbers of T. niger. 1) this urchin is not harvested commercially, 2) it may have several reproductive events per year (Rodriguez and Ojeda, 1993), 3) it has a broad diet (e.g., it feeds on kelp, algal turfs, crastose and drifting algae; Contreras and Castilla, 1987; Rodriguez, 2003) and 4) its main predators (the predatory sea stars H. helianthus and M. gelatinosus), although abundant, only slightly reduce its densities (Gaymer and Himmelman, 2008). The escape responses of T. niger to sea stars, as observed during our study, probably further contribute to its high densities. Further studies are needed to extend our knowledge about the urchin's interaction with its predators, such as tethering experiments and studies using video filming. 25

Predator-prey interactions usually involve co-evolution between the prey and predator: the prey develops mechanisms to improve escaping from predators and the predator develops mechanisms to facilitate the capture of prey. Distance detection of predators by prey, as we show for T. niger, should allow prey to adjust escape behaviours to predatory risk so that wasteful expenditure of energy is limited. This should improve prey survival and fitness. 26

CHAPITRE III

Does the distribution of sea urchins Tetrapygus niger on elevated surfaces represent a strategy for avoiding prédation by sea stars? 27

RÉSUMÉ Nous avons utilisé des expériences dans le milieu naturel pour vérifier si la micro• distribution des oursins de mer Tetrapygus niger sur les surfaces surélevées représente une stratégie permettant de limiter la prédation par les étoiles de mer Heliaster helianthus and Meyenaster gelatinosus. Plusieurs évidences soutiennent cette hypothèse. (1) Un examen de la distribution des oursins et des deux espèces d'étoiles de mer montre que les oursins sont majoritairement situés sur les surfaces surélevées, et les étoiles de mer sur les surfaces plus proches du fond. (2) Lors d'essais impliquant des attaques simulées, le temps nécessaire aux oursins pour briser le contact avec les étoiles de mer était deux fois plus court pour les oursins situés sur les surfaces surélevées que pour ceux situés sur le fond. (3) Lors d'essais impliquant des attaques simulées soutenues (le plus haut niveau de risque de prédation) les oursins pouvaient se détacher pour éviter d'être mangés. Enfin, des expériences au cours desquelles les oursins étaient entravés, indiquent que ceux-ci ont un plus haut taux de survie lorsque situés sur les surfaces surélevées que quand ils sont sur les surfaces plus basses. Nos observations indiquent que M. gelatinosus représente une menace plus forte pour T. niger que H. helianthus. 28

ABSTRACT We ran field experiments to examine if the micro-distribution ofthe sea urchin Tetrapygus niger on elevated surfaces represent a strategy for limiting prédation by the sea stars Heliaster helianthus and Meyenaster gelatinosus. Several lines of evidence supported this hypothesis. (1) A survey of the distribution of the urchin and two sea stars showed that urchins occur manly on elevated surfaces, and sea stars on low surfaces. (2) In trials involving simulated attacks the time needed by the urchin to sever contact with the sea stars was 48 % less on elevated surfaces than on the bottom. (3) In trials involving sustained simulated attacks (high predatory risk) the urchins could detach themselves from the elevated surfaces to avoid being eaten. Finally, tethering experiments indicated that the urchin had a higher survival rate on elevated than low surfaces. Our observations confirm that M. gelatinosus represents a stronger predatory threat to T. niger than H. helianthus. 29

INTRODUCTION Animals have a variety of anti-predatory strategies (Sih et al., 1985; Chivers and Smith, 1998; Ruxton et al., 2004; Caro, 2005) that can be divided in two main categories: (1) avoiding encounters with predators and (2) avoiding being eaten once there has been an encounter (Lima and Dill, 1990). For example, the aposematic coloration ofthe poison frog Dendrobates pumilio decreases the probability of encounters with predators (Saporito et al., 2006) and the autotomizing of arms by the sea star Heliaster helianthus when under attack by the sea star Meyenaster gelatinosus decreases the probability of death (Gaymer and Himmelman, 2008). Predators not only affect prey directly by eating them, but also indirectly by changing their behaviour. Bottom structures, such as crevices, can reduce the probability of encounters with predators and thus increase survival. For several species of sea urchins, the juveniles hide in crevices to reduce the probability of predatory attacks (Scheibling and Hamm, 1991; Rodriguez and Ojeda, 1993; Hereu et al., 2005). Shears and Babcock (2002) show that prédation on the sea urchin Evechinus chloroticus is most intense for juveniles that are beginning to leave crevices for open habitats. Sea urchins are important in benthic communities because of their intensive grazing. Locations supporting high densities of urchins are often transformed into barrens with a reduced diversity and biomass of macroalgae (Himmelman et al., 1983; McClanahan and Shafir, 1990; Alcoverro and Mariani, 2002; Shears and Babcock, 2002). Predators can affect the density, behaviour and population structure of urchins (Tegner and Levin, 1983; Sala et al., 1998; Tuya et al., 2004; Guidetti, 2006). A variety of predators are known to feed on urchins, including sea otters (Estes et al., 1998), fishes (Sala, 1997), lobsters (Andrew and Macdiarmid, 1991), crabs (Scheibling and Hamm, 1991) and sea stars (Himmelman and Dutil, 1991). Relationships between sea stars and urchins are the best studied of these interactions (Jensen, 1966; Rosenthal and Chess, 1972; Dayton et al., 1977; Moitoza and Phillips, 1979; Legault and Himmelman, 1993; Rodriguez and Ojeda, 1998; Hagen et al., 2002). The black sea urchin Tetrapygus niger is the most abundant urchin in central and northern Chile (Vasquez and Buschmann, 1997). Its grazing has converted many subtidal areas into barrens. It usually limits the depth distribution of the subtidal kelp Lessonia trabeculata to shallow water and has caused local extinctions of the subtidal kelp Macrocystis integrifolia (Vega et al., 2005). A number of predators are reported to consume T. niger, including the fishes 30

Semicossyphus maculatus (Fuentes, 1981), Graus nigra (Fuentes, 1982), Pinguipes chilensis (Rodriguez and Ojeda, 1998), Cheilodactylus variegatus, and Oplegnathus insignis (Medina et al., 2004) and the sea star Luidia magellanica (Gaymer and Himmelman, 2008). However, the predators that likely have the greatest impact on the abundance of T. niger are the sea stars H helianthus and M. gelatinosus (Barrios et al., 2008; Gaymer and Himmelman, 2008). These sea stars have been described as keystone predators in shallow rocky communities in northern Chile; (Gaymer and Himmelman, 2008; Barahona and Navarrete, 2010). H helianthus is a generalist feeder consuming prey according to their availability (Gaymer and Himmelman, 2008; Barahona and Navarrete, 2010), whereas M. gelatinosus is a selective feeder that prefers consuming the urchin T. niger (Gaymer and Himmelman, 2008). Both H. helianthus and M. gelatinosus are endemic along the west coast of South America and in northern Chile their average densities are 3 ind. m2 and <0.5 ind. m2, respectively (Dayton et al., 1977; Tokeshi et al., 1989; Gaymer and Himmelman, 2008; Navarrete and Manzur, 2008). H. helianthus has a robust and flattened body with as many as 40 arms and can attain up to 32 cm in diameter (Tokeshi et al., 1989). In contrast, M. gelatinosus has a soft body with 6 thick arms and can attain 56 cm in diameter (Dayton et al., 1977; Gaymer and Himmelman, 2008). Three studies have reported the behavioural responses of T. niger to its sea star predators. Dayton et al. (1977) mentioned that T. niger can detect M. gelatinosus at a distance and indicated that it responds rapidly by fleeing. Rodriguez and Ojeda (1998) examined the urchin's responses to M. gelatinosus and the fish Pinguipes chilensis in the laboratory and reported that the urchin increases its rate of displacement when these predators were nearby and that a higher proportion of individuals respond in trials with M. gelatinosus (92 %) than with P. chilensis (67 %). Finally, Urriago et al. (Chapter II) quantified the urchin responses to predatory sea stars, and showed that the urchin can differentiate between predatory (H. helianthus and M. gelatinosus) and non- predatory (Stichaster striatus) sea stars, can distinguish between different threat levels associated with the predatory sea stars and can detect predatory sea stars at a distance. They indicated that M. gelatinosus presents a stronger predatory threat to urchins than H. helianthus. T niger frequently occur on elevated surfaces (e.g., boulder tops) within the barrens communities that predominate in shallow rocky subtidal areas along the coast of central and northern Chile. As food resources are less abundant on elevated surfaces than on the bottom, we reasoned that this micro-distribution of T. niger might limit attacks by the sea stars H. helianthus 31

and M. gelatinosus. The present study examines this hypothesis. We first documented the preference of the urchins for elevated surfaces, then examined the urchin's responses to sea stars in a variety of situations (predatory attacks), and finally conducted a tethering experiment to compare the survival on high and low surfaces (i.e. top and bottom of boulders).

METHODS Our study was conducted during June, July and August in 2008 and 2009 in the subtidal zone at Obispito Bay (26°48'22"S, 70°47*05"W), Cisnes Bay (27°14'50"S, 70°57'34"W) and El Frances Bay (30°5'42"S, 71°22'47"W) in northern Chile. All manipulations were made using SCUBA diving at depths of 2 and 9 m in wave-exposed environments (i.e., with continuous back and forth water movement). At the three bays, the bottom was moderately sloped (down to -10 m depth at -30 m from shore) and supported a barrens community, with a scarcity of fleshy macroalgae. The sea urchin Tetrapygus niger was abundant and the sea stars Heliaster helianthus and Meyenaster gelatinosus were present in much lower numbers. During our trials, water temperatures ranged between 12 and 14 °C In the various experiments, the sea stars were taken at random from among individuals that were stationary and not feeding. Different urchins and sea stars were used in each trial.

Distribution of sea urchins and sea stars We made a field survey at Obispito, Cisnes and El Frances Bays to characterized the abundance and distribution of T. niger and the two predatory sea stars H helianthus and M. gelatinosus. For each urchin and each sea star encountered we recorded its position in two categories, high and low surfaces. We also quantified the percentage cover of high and low surfaces. High surfaces included from the tops of boulders and bedrock outcrops to half way down the vertical faces ofthe structures, and low surfaces were mainly flat areas of pebbles, shell debris and small cobbles but also included surfaces extending half way up the sides of boulders and outcrops. The boulders were 1.0-1.5 m in height and the outcrops 4-5 m. In 2008, we systematically surveyed an entire cove at Cisnes Bay (using five 50-m transects running from the shore seaward, and spaced at 6-m intervals) whereas in 2009 we sampled 79 randomly placed 1 - m2 quadrats at Obispito Bay and 72 at El Frances Bay. 32

We made trials in which we placed predatory sea stars on boulder tops to determine whether they would remain there or move to lower positions. In each trial we placed a sea star on a boulder top not covered by sea urchins and held it there until it was attached (<1 min). Then, after 5 min we recorded its position. We ran 20 trials for both M. gelatinosus and H. helianthus at El Frances Bay in 2009.

Responses of sea urchins to sea stars on different types of bottom We further performed a number of short-term field experiments at Cisnes Bay in 2009 to provide insights into the responses of the sea urchin to the two predatory sea stars, H. helianthus (9-14 cm in radius) and M. gelatinosus (13-21 cm), on three types of bottom. We first quantified the time it took isolated urchins (5-6 cm in diameter and at least 20 cm from other urchins) to sever contact from a simulated attack by a sea star. A simulated attack consisted of holding a sea star so that an arm, or several arms in the case of H. helianthus, covered half of the target urchin. The sea star was maintained in the initial position, even as the urchin moved away. In each trial on each type of bottom we first placed an urchin on the substratum and allowed it 4-5 min to attach (this was done because urchins were rarely found on irregular horizontal surfaces). The urchins always remained very close to where they were placed. Then we initiated the simulated attack. We executed 20 simulated attacks with both H. helianthus and M. gelatinosus (1) on irregular horizontal bottoms, (2) on relatively smooth horizontal platform (bedrock with few surface irregularities), and (3) on relatively smooth vertical walls (side of a bedrock outcrop or large boulder).

Sustained (simulated) attacks on vertical walls We further ran experiments at Cisnes Bay in 2008 to examine the responses of undisturbed sea urchins to a sustained simulated attack (hereafter referred to as "sustained attack") by predatory sea stars on vertical walls. In each trial we first selected a target urchin at the lower edge of an aggregation at the top of a wall. The walls were 4-5 m in height and the distance between the target urchin and the aggregation at the top varied from 50 to 80 cm. We then initiated a sustained attack from below the urchin and advanced the sea star so that its arm (or several arms with H. helianthus) always covered half the urchin. Each trial was continued 33 until the urchin detached or reached the aggregation at the top of the wall. We executed 20 trials with both H. helianthus and M. gelatinosus.

Sustained attacks on aggregations We also examined the behaviour of aggregations of urchins (5-23 individuals) to sustained attacks by predatory sea stars on boulder tops. In each trial we held a sea star so that it covered about half of an urchin at the edge ofthe aggregation. If this urchin moved, we continued to hold the sea star over it, and if it detached, we continued by attacking the next urchin in the same way. In all trials the aggregation started moving away from the sea star shortly after we initiated the first sustained attack. We continued with this procedure until all the urchins had (1) detached, (2) moved half-way down and then to the side (around the boulder), or (3) moved to the bottom. When no urchins were on top of the boulder, we removed the sea star. Then after 3 min we recorded the numbers of urchins that (1) continued to flee, (2) climbed up a boulder (the same boulder or a nearby boulder), or (3) remained stationary on the bottom. Preliminary trials show that 3 min was long enough for an urchin to return to a boulder top. We executed 15 trials with both M. gelatinosus and H. helianthus at Cisnes Bay in 2009. Each trial was on a different boulder and with a different sea star.

Survival on high and low surfaces In 2008 at Cisnes Bay we performed trials with tethered sea urchins on boulder tops, and on the bottom around boulders, to estimate the probability of survival in the two contrasting positions. The state ofthe each tethered sea urchin was assessed after 24 h, and when possible the predators causing mortalities were identified, from direct observations or from prey remains. In both high and low positions, we ran 11 trials with small urchins (15-20 mm) and 11 with large urchins (50-60 mm). The urchins were attached with monofilament threads to the center of a 20 x 20 cm ceramic tile (6 holes were drilled through the tiles to facilitate attaching the urchins). In the trials in high positions the tiles were attached to the boulders with "sea-goin" epoxy resin, whereas in the trials at low positions, rocks were placed at the edge of the tiles to hold them on the bottom. 34

Statistical analyses To evaluate if distribution of the sea urchin T. niger and the two sea stars H. helianthus and M. gelatinosus on high and low surfaces corresponded to the availability of these surfaces, •y we used two procedures. We used a x test to evaluate the data from the systematic survey at Cisnes Bay, and the SURVEYMEANS procedure to evaluate the data from the randomly sampled quadrats at Obispito and El Frances Bays. We used a two-way ANOVA to compare the mean time that urchins took to sever contact with sea stars in the trials on the three types of bottom. In this model there were two fixed factors, the sea star species (H. helianthus and M. gelatinosus) and the type of bottom (irregular horizontal, smooth horizontal and vertical walls). The data were log transformed to meet the assumptions of normality and homogeneity of variance. Pairwise comparisons were performed using protected Fisher least square difference tests (LSD). For the experiments evaluating the behavioural responses of urchins aggregated on boulder tops, we analyzed the proportion of urchins (1) detaching, (2) moving down and then to the side and (3) moving to the bottom, in response to sustained attacks by a sea star, then, 3-min after the sea star was removed, the proportion (1) continuing to flee, (2) climbing up a boulder and (3) remaining stationary on the bottom. The same model was used for both analyses. These responses were estimated from odds ratios calculated with a multinomial logistic regression model (using the LOGISTIC procedure and the glogit link). Finally, a generalized linear model (using the GENMOD procedure with the binomial distribution and the logit link) was used to compare the proportion of small and large tethered urchins that survived on high and low surfaces. In this model there were two fixed factors, the position (high and low) and the size of the urchins (small and large). All statistical test were performed with the software SAS 9.2 (2008).

RESULTS Distribution of sea urchins and sea stars There was a similar availability of high and low surfaces at Cisnes Bay (about 50 % for each), whereas high surfaces predominated at Obispito Bay (81 %) and El Frances Bay (72 %). At all three bays, most sea urchins T niger were found on high surfaces (98 % at Cisnes, 99 % at Obispito and 67 % at El Frances; PO.0001). In contrast, the predatory sea stars H. helianthus 35 and M. gelatinosus were more common on low surfaces (PO.0001), except for H helianthus at Obispito Bay where only 33 % were found on low surfaces (P=0.20). The scarcity of low surfaces at Obispito Bay (only 20 %) probably contributed to the high frequency of//, helianthus on high surfaces at this location. In most cases the proportions of urchins and of the two predatory sea stars on high and low surfaces did not correspond to the availability of these surfaces (Fig. 3.1). The exceptions were urchins at El Frances Bay and H. helianthus at Obispito Bay, where proportions corresponded to availability. In the experiments in which we placed predatory sea stars on boulder tops (without urchins) and observed their positions 5 min later, 50 % of the H helianthus remained on the boulder tops and 50 % moved to the bottom. In contrast, in the parallel trials with M. gelatinosus only 5 % stayed on the boulder tops and 95 % moved to the bottom.

Responses of sea urchins to sea stars on different types of bottom On all three types of bottoms, the urchins first responded to simulated attacks by raising the spines and then moved away in the opposite direction to the point of contact with the predator (these responses are described in greater detail in Chapter II). The time for the urchins to sever contact with the sea stars did not vary with the sea star predator, H helianthus or M. gelatinosus, but varied with the type of bottom (PO.0001; Fig. 3.2). Contact was severed most rapidly in the trials on smooth horizontal platform (24 s), least rapidly on irregular horizontal bottoms (46 s), and at an intermediate level on relatively smooth vertical walls (36 s).

Sustained (simulated) attacks on vertical walls In the trials examining the proportion of urchins detaching when subjected to sustained attacks by predatory sea stars on relatively smooth vertical walls, the target urchin first moved up the wall to distance itself from the sea star. In trials with H helianthus, 30 % of the urchins detached from the vertical wall before reaching the urchin aggregation at the top. This compared to 75 % of the urchins in the trials with M. gelatinosus. The targeted urchins that did not detach moved to the edge of the aggregation and almost immediately tried to climb over it. They failed but then moved around the aggregation. Almost all of the urchins in the aggregation at the top moved away a few seconds after the targeted urchin came into contact with the aggregation. 36

Cisnes Bay

Availability of high surfaces

T. niger H. helianthus M. gelatinosus

Figure 3.1. Proportion (%) of the sea urchin Tetrapygus niger and the predatory sea stars Heliaster helianthus and Meyenaster gelatinosus on high and low surfaces relative to the availability of these surfaces at Obispito, Cisnes and El Frances Bays. Values are means ± SE; the entire cove was surveyed at Cisnes Bay (thus we do not show SEs), whereas quadrats were sampled at the other bays. An asterisk (*) indicates a significant difference between the proportion of surfaces used and that available (PO.0001). 37

Simulated attacks on different «types of bottoms H. helianthus ■ M. gelatinosus

Irregular horizontal Smooth horizontal Smooth vertical

Figure 3.2. Mean time for the sea urchin Tetrapygus niger to sever contact with the two predatory sea stars, Heliaster helianthus and Meyenaster gelatinosus, on irregular horizontal bottom, smooth horizontal bottom and smooth vertical walls. Values are means ± SE. Treatments not sharing the same letter are different (PO.0001). In all treatments, the response did not vary between the two sea stars (P=0.30). 38

Sustained attacks on aggregations The urchin aggregations on boulder tops showed strong escape responses when subjected to sustained attacks by H. helianthus and M. gelatinosus at the edge of urchin aggregations, and the proportions of urchins in the different behavioural categories (out of the total number of urchins on the boulder tops) did not vary with the sea star species (Fig. 3.3, P=0.16). In the trials with both two sea stars, all targeted urchins fled in the opposite direction and soon after detached. These represented 82 % of the urchins in the trials with H. helianthus and 72 % in the trials with M. gelatinosus. Nearby urchins that did not have contact with the sea stars did not detach but moved away and eventually either (1) moved down and then to the opposite side ofthe boulder or (2) moved all the way down to the bottom. The proportions of urchins in the latter two categories did not vary between trials with the two sea stars (P=0.21, Fig. 3.3). Following the removal of the predatory sea stars, most of the urchins (-80 %, including individuals that had moved to the sides of the boulders during the attacks) climbed back up the sides of the experimental boulder or an adjacent boulder. Only about 18 % continued to flee and about 2 % remained stationary (Fig. 3.3). These proportions did not vary between trials with the two sea stars (P=0.62). 39

Sustained (simulated) attacks

Detached " n=20

Moved down and to the side H. helianthus

M. gelatinosus Moved to the bottom

3 min after the sea star was removed

Continued to flee

Climbed up the boulder

Remained stationary

—i— 20 40 60 80 100 Proportion (%)

Figure 3.3. Proportion (%) of urchins Tetrapygus niger in aggregations on boulder tops that detached, moved down and to the side and moved to the bottom in response to sustained attacks by the predatory sea stars Heliaster helianthus and Meyenaster gelatinosus and then, 3 min after the predatory sea star was removed, the proportion of urchins that continued to flee, climbed up a boulder and remained stationary on the bottom. Treatments not sharing the same letter are different (PO.0001). The sea star species did not affect how the urchins responded to sustained attacks (P=0.16) or the response ofthe urchins after removal ofthe sea stars (P=0.62). 40

Survival on high and low surfaces The 24-h trials with tethered sea urchins indicated that survival was 36 % higher on boulder tops than on the bottom for both small and large urchins (P=0.01; Fig. 3.4). Although the proportion of individuals lost tended to be higher for small than large urchins, the difference was not significant (P=0.11). When we returned at the end ofthe 24-h trials, we observed four types of predators attacking the urchins: (1) the sea stars H. helianthus and M. gelatinosus, (2) the rockfish Scartichthys viridis, (3) the gastropods Tegula atra and Crassilabrum crassilabrum, and (4) the hermit crabs Pagurus spp (Fig. 3.5). The rockfish was not observed killing tethered urchins but was seen biting off podia. All attacks by sea stars were by single individuals that everted their stomach over the urchin, whereas the gastropods and the hermit crabs always attacked in groups, and sometimes both snails and hermit crabs attacked at the same time (they aggregated around the edge of the urchin). We were unable to identify the predators causing many of attacks on small urchins (labeled unknown in Fig. 3.4), because the attacks were rapid and there were no prey remains (no urchins escaped in control trials in enclosures). After the 24-h trials, a number of the urchins were observed with missing spines (the scars appeared as white spots) that we suspect were the result offish attacks. 41

.KO Top D M. gelatinosus 70 ■ H. helianthus

■ Tegula sp. n=n 60 B Pagurus spp.

50 13 Crassilabrum sp. □ Unknown 40

30

20

10

0

80 - Bottom o 2 70 - 60 -

50 - &&&&&& 40 - 30 - ■ 20 - wmmm

10 -

0 - i Small Large

Figure 3.4. Rate of mortality of sea urchins Tetrapygus niger in experiments where small (15-20 mm) and large (50-60 mm) urchins were tethered to boulder tops and to the bottom for 24-h. The different shades identify the predators that were observed. Mortality did not vary with urchin size (small and large, P=0.11), but did with position (bolder tops and bottom; P=0.01). 42

Figure 3.5. Predatory attacks on tethered sea urchins Tetrapygus niger by (a) the sea star Heliaster helianthus; (b) the sea star Meyenaster gelatinosus; (c) the gastropods Tegula atra and Crassilabrum crassilabrum, and hermit crabs Pagurus spp.; and (d) an urchin showing loss of aboral spines, probably due to attacks by the rockfish Scartichthys viridis. 43

DISCUSSION Our quantitative survey in three bays showed that the sea urchin Tetrapygus niger most occurs on elevated surfaces. Almost all the urchins at Cisnes Bay (98 %) and Obispito Bay (99 %) and a large proportion of urchins at El Frances Bay (67 %) were found on elevated surfaces. During previous studies of barrens communities, where the substratum was boulders and bedrock outcrops, we also noted the association of urchins with elevated surfaces. Our study provided several lines of evidence indicating that urchins on elevated surfaces are exposed to less risk of attacks by predators. First, elevated surfaces correspond to habitats where the important urchin predators, the sea stars Heliaster helianthus and Meyenaster gelatinosus, occur in reduced numbers. For example, our survey at the three bays showed that these two sea stars were predominantly (>75 %) found on lower surfaces, although there was an exception for H helianthus at Obispito Bay, where low surfaces were rare. The preference of sea stars, particularly M. gelatinosus, for bottom surfaces may be related to the risk of being detached by wave action. H helianthus may also prefer bottom areas because its prey species, other than the urchins, are more available there. Sea stars are also more likely to fall from elevated surfaces when attacking urchins, as an attacking sea star is less securely attached and its raised arms are more exposed to waves. The sensitivity of H helianthus to wave action was noted by Barahona and Navarrete (2010) as they observed that its movement in the intertidal zone was reduced when there was increased wave action. Further, Gaymer et al. (unpubl. data) found that H helianthus is more abundant in sheltered than in exposed areas. Although both M. gelatinosus and //. helianthus are generally most abundant in bottom habitats, H. helianthus tends to use elevated surfaces more than M. gelatinosus. For example, in the trials in which we placed the two sea stars on boulder tops, only 50 % ofthe H. helianthus moved off the boulder tops in 5 min, compared to 95 % for M. gelatinosus. H. helianthus may be better adapted to wave-swept areas than M. gelatinosus because of its many arms and more flattened body (Gaymer and Himmelman, 2008). The high frequency of T. niger on elevated surfaces suggests that it is well adapted to wave activity. We have further noted that it remains on elevated surfaces even during periods of increased wave activity. The urchin usually keeps its spines lowered giving it a flattened profile that should limit the impact of waves. Ryer et al. (2004) similarly reports that the flatfish Hippoglossus stenolepis prefers high surfaces (emergent structures such as sponges) and suggests that its swimming among these structures protects it from prédation. 44

A second factor that may lead T. niger to prefer to aggregate on elevated surfaces is that there are usually fewer irregularities in the substratum than on bottom surfaces (JDU, observations). Irregularities (caused by pebbles, shell debris and small cobbles as well as crevices) can block the urchin from fleeing from sea stars. For example, in our trials with simulated attacks on individuals, the urchins needed almost twice as much time to sever contact with a sea star on an irregular horizontal bottom than on a smooth horizontal bottom. Sea stars readily move over bottom irregularities because of their large size and extended arms, whereas urchins must follow the contours of small depressions and ridges. On three occasions during dives we observed an urchin that was fleeing from M. gelatinosus that became trapped in a crevice. In each instance the sea star climbed onto the urchin, wrapping its arms around it, and then everted its cardiac stomach to begin feeding. Our diving observations of interactions between urchins and sea stars suggest that it is almost impossible for H. helianthus and M. gelatinosus to catch T. niger on smooth surfaces where its movement is unimpeded. Similarly Dayton et al. (1977) observed that the urchin Loxechinus albus must be blocked by bottom irregularities for attacks of M. gelatinosus to be successful. In our trials involving sustained attacks of urchin aggregations on boulder tops, most urchins (that had detached or moved down to the bottom) climbed up the sides of the boulders once the sea stars were removed. The return to the boulder tops in the absence of the sea star suggests that this behaviour in and of itself would limit attacks by sea stars on lower surfaces. Phillips (1975; 1976) found that the gastropods Acmaea spp. climbed vertical surfaces when they detected the odour (without contact) of predatory sea stars (Pisaster ochraceus, Pisaster giganteus, Pycnopodia helianthoides, and Leptasterias aequalis). This contrasts with T niger which moves to elevated surfaces even when there is no predatory stimulus. A third advantage of elevated surfaces is that they provide the urchin with the alternative of detaching from the substratum ("jumping ship") to avoid being eaten. This was seen when we made sustained attacks (1) on individual urchins on vertical walls and (2) on urchin aggregations on boulder tops. Detachment caused the urchins to fall and be transported by wave surge, thus rapidly distancing them from the sea stars. Although "jumping ship" prevents being eaten, it entails costs. Spines are likely broken, so that energy must be expended for spine repair (Ebert, 1968; Edwards and Ebert, 1991). Further, urchins usually land on the bottom "oral-side-up" and must extend podia to right themselves. During our experiments we often observed the rockfish S. 45

viridis biting the podia of urchins that were righting themselves after falling from walls. If an urchin were unlucky enough to fall near another sea star, it would have to right itself before it was able to flee. Our data suggest that urchins only resort to detaching when the risk of prédation is extreme. This was indicated because the urchins simply fled in response to being subjected to a simulated attack, whereas they almost always detached when subjected to a sustained simulated attack (trials on vertical walls and boulder tops). Other prey species have also been reported to detach from the bottom to avoid predatory attacks. For example, laboratory studies by Alexander and Covich (1991) show that the gastropod Physella virgata detaches from the bottom and floats upward when touched by predatory crayfish. When H helianthus encounters M. gelatinosus, which is also its predator, it suddenly raises its arms ("crown position"). This involves detachment from the bottom, making it more likely to be transported by wave action (Gaymer and Himmelman, 2008). Snyder and Snyder (1970) provides another example on an animal detaching in response to a perceived risk. They exposed urchins Diadema antillarum under a strong unidirectional flow to odours of injured conspecifics and observed that individuals lost hold on the substratum and were carried by the current. Finally, diving observations by Dayton et al. (1977) showed that various prey species detach and fall to the bottom when they detect the odour of M gelatinosus. Thus, detachment and use of water flow may be a common strategy used to limit predatory attacks. Our trials with tethered urchins provided strong evidence that survival is greater on high than low surfaces but there was no statistical difference in survival between small (15-20 mm) and large urchins (50-60 mm). About half of the predators that could be identified during these trials were sea stars. In contrast, in laboratory trials Rodriguez and Ojeda (1998) reported that M. gelatinosus pass over small urchins (20-30 mm) without causing damage, preferring to attack large urchins (35-60 mm). Observations of prey-size selection of feeding sea stars encountered during an extensive benthic survey (CF. Gaymer, unpubl. data) showed that the most common size group of urchins, individuals measuring 40-60 mm in diameter, was eaten 62 % of the time by M. gelatinosus and 65 % of the time for H helianthus. In their survey the sizes of urchins consumed by M. gelatinosus closely reflected availability, whereas H. helianthus tended to select smaller urchins. Although tethering undoubtedly reduces the ability of urchins to escape from predators, this technique nevertheless provides a tool for identifying predators and comparing levels of prédation in different habitats (Shears and Babcock, 2002; Guidetti and Dulcic, 2007). 46

We observed a variety of animals feeding on T niger that have not previously been reported to prey on urchins. These included the rockfish Scartichthys viridis, the gastropods Tegula atra and Crassilabrum crassilabrum, and hermit crabs Pagurus spp. Our study is the first to report that T atra includes animals in its diet. M. gelatinosus likely represents a greater predatory threat to T niger than H helianthus. This was indicated because the urchins submitted to sustained attacks (extreme prédation risk) on vertical walls detached 45% more often when the sea star predator was M. gelatinosus than in when it was H helianthus. Further, M. gelatinosus preferentially feeds on the urchin (Gaymer and Himmelman, 2008). In contrast H. helianthus is a generalist feeder that most frequently consumes mussels, barnacles and small gastropods (Barahona and Navarrete, 2010). In a parallel study on responses of T. niger to the two sea stars (Chapter II), we provide further evidence that M. gelatinosus represents a stronger threat than H helianthus. We showed that the urchin severed contact faster with M. gelatinosus than with H. helianthus when subjected to simulated attacks, and that the urchin detects M. gelatinosus at greater distances than H helianthus. Many studies report that the distribution of urchins is related to the distribution of their foods, for example urchins often form grazing fronts at the edge of kelp beds or aggregate on algal debris that had been carried by currents to urchin barrens (Vadas et al., 1986; Himmelman and Nedelec, 1990; Scheibling and Hamm, 1991; Hagen and Mann, 1994; Vadas and Elner, 2003). Field observations by Rodriguez and Farina (2001) similarly report that T. niger aggregates on drift kelp Macrocystis pyrifera, and Rodriguez and Ojeda (1998) found that T niger aggregated on drift algae Lessonia sp. added to experimental tanks. It is unlikely that the micro-distribution of T niger on elevated surfaces is related to food resources because food is rare on elevated surfaces. Boulder tops are devoid of fleshy algae and any macroalgal debris present usually sinks to depressions in the bottom. Thus, the urchin T niger aggregates on elevated surfaces and moves to higher surfaces even when not being pursued by a sea star. Our study indicates that this choice of microhabitat represents an adaptation for avoiding being eaten by two common predatory sea stars. Occupying elevated surfaces might increase exposure to pelagic predators such as fish. A number of studies involving tethering experiments report that prédation on other species of urchins (small to medium sizes) is higher at locations where the density of predatory fish is high than in areas where it is low (Sala and Zabala, 1996; Shears and Babcock, 2002; Guidetti, 2006). In Chile, the 47 abundance of fish that feed on T niger is presently low because of overfishing (Godoy et al., 2010). Thus, the decrease in fish predators may have led to an increase in the importance ofthe sea stars H. helianthus and M. gelatinosus as predators of urchins. Bonaviri et al. (2009) similarly suggest that a decrease in numbers of predatory fish (Diplodus sargus and D. vulgaris) may have increased the importance of the sea star Marthasterias glacialis in controlling the abundance of the urchins in the Mediterranean Sea. Additional studies are needed to further understand the effects of predators on the distribution and abundance of T niger. It would be particularly useful to use 24-h video of tethered urchins to document the types of predators that attack urchins during different periods of the day. Such studies would at the same time indicate whether there are predators that have not hitherto been observed. 48

CHAPITRE IV Conclusion générale Cette étude est la première à examiner les réponses de l'oursin de mer Tetrapygus niger envers ces prédateurs sous des conditions de courants bidirectionnels induits par les vagues. Mes expériences montrent que les oursins ont développé de nombreuses réponses pour éviter la prédation par les étoiles de mer Heliaster helianthus and Meyenaster gelatinosus. Lors d'essais utilisant des attaques simulées j'ai démontré que T. niger fait clairement la différence entre les étoiles de mer prédatrices (H. helianthus et M. gelatinosus) et non-prédatrices (Stichaster striatus et une fausse étoile de mer). En réponse aux attaques simulées par les étoiles prédatrices H. helianthus et M. gelatinosus, T. niger hérisse presque immédiatement ses épines à la verticale. Puis, il déplace et étire ses pieds ambulacraires. Ces deux comportements se produisent pratiquement au même moment. Finalement, le mouvement de déplacement (fuite) amène l'oursin à perdre le contact avec l'étoile de mer. L'étoile de mer chilienne Luidia magellanica se nourrissant également de T. niger, il serait intéressant d'évaluer les réponses induites par ce prédateur. J'ai démontré la capacité de l'oursin à faire la distinction entre différents niveaux de risque de prédation par les étoiles de mer prédatrices ainsi que l'augmentation de la réponse de fuite avec ce même risque. Lors de mon étude, la situation de risque maximum était une attaque continue et soutenue au cours de laquelle une étoile de mer était maintenue sur un oursin y compris lors de sa fuite. Le niveau de risque suivant était une attaque simulée, au cours de laquelle une étoile de mer était maintenue en une position et ne pouvait donc pas suivre l'oursin, puis j'ai utilisé un simple contact de l'extrémité d'un bras d'étoile de mer. Enfin, j'ai fait des tests au cours desquels les étoiles de mer étaient placées à différentes distances des oursins. Le niveau de risque testé le plus bas correspondait à une étoile de mer placée à 50 cm d'un oursin. Les saponines pourraient être impliquées dans les phénomènes de chimiodétection à distance chez l'oursin. D'intéressantes études futures pourraient viser à l'identification des différents types de saponines sécrétés par H. helianthus et M. gelatinosus, ainsi qu'à la quantification de la réponse des oursins à des concentrations spécifiques de ces composés chimiques. Mes études ont clairement démontré l'existence de la chimiodétection à distance chez T. niger dans des conditions naturelles de houle (vagues). La détection se faisait sur de plus grandes distances dans le cas de M. gelatinosus (les oursins continuaient de répondre à une distance de 50 49

cm) que dans le cas de H. helianthus (pas de réponse à partir de 30 cm). D'autres observations indiquaient également que T. niger associait un niveau de prédation plus élevé à la présence de M. gelatinosus qu'à celle de H. helianthus. Lors des attaques continues et soutenues (le plus haut niveau de risque de prédation) sur des murs verticaux, les oursins se détachent de la surface 45% plus souvent avec M. gelatinosus qu'avec H. helianthus. Finalement, lors des essais au cours desquels les étoiles de mer étaient ajoutées au milieu d'agrégats d'oursins, ceux-ci s'enfuyaient plus rapidement et sur de plus grandes distances en réponse à la présence de M. gelatinosus qu'en présence de H. helianthus. Je n'ai pas trouvé d'effets de la densité de la population d'oursin sur la manière dont les agrégats répondaient à la prédation des étoiles de mer. Mon étude dans trois baies (sites naturels présentant une abondance de rochers) a montré que les oursins sont majoritairement trouvés en agrégats serrés sur le sommet des rochers (boulders). Quelques fois ces agrégats s'étendent jusqu'à la moitié de la hauteur sur les cotés des rochers. Nos observations suggèrent que la micro-distribution de T. niger sur les surfaces surélevées (dans les zones dénudées dominées par la présence de rochers et de roche mère) représente une stratégie visant à réduire la prédation par les étoiles de mer. J'ai produit plusieurs observations qui supportent cette hypothèse: (1) le nombre d'étoiles de mer prédatrice (H. helianthus and M. gelatinosus) est inférieur sur les surfaces surélevées par rapport au fond, (2) l'action des vagues sur les surfaces surélevées rend l'attaque des oursins par les étoiles de mer plus difficile. (3) Les irrégularités du substrat (qui facilitent la capture des oursins) sont plus rares sur les surfaces surélevées, et (4) les surfaces surélevées permettent aux oursins de se détacher pour se distancer volontairement des attaques de leurs prédateurs. Pour poursuivre les études, je suggère de procéder à des études expérimentales utilisant le marquage des individus pour évaluer le taux de mouvement des oursins entre les surfaces surélevées et le fond (e.g., pour combien de temps les oursins restent-ils sur le haut des rochers? A quelle fréquence descendent-ils sur le fond?), de paire avec des traitements impliquant la présence ou l'absence des étoiles de mer prédatrices. Finalement, mes expériences impliquant l'entravement des oursins ont montré que ceux-ci ont un taux de survie plus élevé sur les surfaces élevées qu'au fond. Il serait pertinent de filmer les oursins entravés sur des périodes de 24 h de façon à mieux documenter quel type de prédateur attaque les oursins à quel moment de la journée. De telles études pourraient (en même temps) révéler l'existence de prédateurs n'ayant pas encore été identifiés. 50

Références bibliographiques

Alcoverro, T., Mariani, S., 2002. Effects of sea urchin grazing on seagrass (Thalassodendron ciliatum) beds of a Kenyan lagoon. Mar. Ecol. Prog. Ser. 226, 255-263. Alexander, J.E., Covich, A.P., 1991. Predator avoidance by the freshwater snail Physella virgata in response to the crayfish Procambarus simulons. Oecologia 87,435-442. Andrew, N.L., MacDiarmid, A.B., 1991. Interrelations between sea urchins and spiny lobsters in northeastern New Zealand. Mar. Ecol. Prog. Ser. 70, 211-222. Barahona, M., Navarrete, S.A., 2010. Movement patterns ofthe seastar Heliaster helianthus in central Chile: relationship with environmental conditions and prey availability. Mar. Biol. 157,647-661. Barrios, J.V., Gaymer, CF., Vasquez, J.A., Brokordt, K.B., 2008. Effect of the degree of autotomy on feeding, growth, and reproductive capacity in the multi-armed sea star Heliaster helianthus. J. Exp. Mar. Biol. Ecol. 361, 21-27. Bernstein, B.B., Williams, B.E., Mann, K.H., 1981. The role of behavioral responses to predators in modifying urchins' (Strongylocentrotus droebachiensis) destructive grazing and seasonal foraging patterns. Mar. Biol. 63, 39-49. Bernstein, B.B., Schroeter, S.C, Mann, K.H., 1983. Sea urchin (Strongylocentrotus droebachiensis) aggregating behavior investigated by a subtidal multifactorial experiment. Can. J. Fish. Aquat. Sci. 40, 1975-1986. Bonaviri, C, Fernandez, T.V., Badalamenti, F., Gianguzza, P., Di Lorenzo, M., Riggio, S., 2009. Fish versus prédation in controlling sea urchin populations in Mediterranean rocky shores. Mar. Ecol. Prog. Ser. 382, 129-138. Broom, M., Higginson, A.D., Ruxton, G.D., 2010. Optimal investment across different aspects of anti-predator defences. J. Theor. Biol. 263, 579-586. Brown, G.E., Godin, J.G.J., 1997. Anti-predator responses to conspecific and heterospecific skin extracts by threespine sticklebacks: alarm phéromones revisited. Behaviour 134, 1123- 1134. Campbell, A.C, Coppard, S., D'Abreo, C, Tudor-Thomas, R., 2001. Escape and aggregation responses of three to conspecific stimuli. Biol. Bull. 201, 175-185. Caro, T., 2005. Antipredator defenses in birds and mammals. Chicago University Press, Chicago. Chivers, D.P., Smith, R.J.F., 1998. Chemical alarm signalling in aquatic predator-prey systems: a review and prospectus. Ecoscience 5, 338-352. Contreras, S., Castilla, J.C, 1987. Feeding behavior and morphological adaptations in two sympatrie sea urchin species in central Chile. Mar. Ecol. Prog. Ser. 38, 217-224. Covich, A.P., Crawl, T.A., Alexander, J.E., Vaughn, CC, 1994. Predator-avoidance responses in freshwater decapod-gastropod interactions mediated by chemical stimuli. J. N. Am. Benthol. Soc. 13,283-290. Cronin, G., Hay, M.E., 1996. Induction of seaweed chemical defenses by amphipod grazing. Ecology 77, 2287-2301. Crawl, T.A., Covich, A.P., 1990. Predator-induced life-history shifts in a fresh-water snail. Science 247, 949-951. Dale, J., 1997. Chemosensory search behavior in the starfish Asterias forbesi. Biol. Bull. 193, 210-212. Dayton, P.K., 1975. Experimental evaluation of ecological dominance in a rocky intertidal algal community. Ecol. Monogr. 45, 137-159. 51

Dayton, P.K., Rosenthal, R.J., Mahen, L.C, Antezana, T., 1977. Population structure and foraging biology of predaceous Chilean asteroid Meyenaster gelatinosus and escape biology of its prey. Mar. Biol. 39, 361-370. Drolet, D., Himmelman, J.H., 2004. Role of current and prey odour in the displacement behaviour ofthe sea star Asterias vulgaris. Can. J. Zool. 82, 1547-1553. Ebert, T.A., 1968. Growth rates ofthe sea urchin Strongylocentrotus purpuratus related to food availability and spine abrasion. Ecology 49, 1075-1091. Edwards, P.B., Ebert, T.A., 1991. Plastic responses to limited food availability and spine damage in the sea urchin Strongylocentrotus purpuratus (Stimpson). J. Exp. Mar. Biol. Ecol. 145, 205-220. Espoz, C, Castilla, J.C, 2000. Escape responses of four Chilean intertidal limpets to seastars. Mar. Biol. 137, 887-892. Estes, J.A., Tinker, M.T., Williams, T.M., Doak, D.F., 1998. Killer whale prédation on sea otters linking oceanic and nearshore ecosystems. Science 282, 473-476. Feder, H.M., 1963. Gastropod defensive responses and their effectiveness in reducing prédation by . Ecology 44, 505-512. Feder, H.M., 1972. Escape responses in marine invertebrates. Sci. Amer. 227, 93-100. Fishlyn, D.A., Phillips, D.W., 1980. Chemical camouflaging and behavioral defenses against a predatory seastar by 3 species of gastropods from the surfgrass Phyllospadix community. Biol. Bull. 158, 34-48. Fuentes, H.R., 1981. Feeding habit of Semicossyphus maculatus (Labridae) in coastal waters of Iquique in northern Chile. Jap. J. Ichthyol. 27, 309-315. Fuentes, H.R., 1982. Feeding habits of Graus nigra (Labridae) in coastal waters of Iquique in northern Chile. Jap. J. Ichthyol. 29, 95-98. Furrow, F.B., Amsler, CD., McClintock, J.B., Baker, B.J., 2003. Surface sequestration of chemical feeding deterrents in the Antarctic sponge Latrunculia apicalis as an optimal defense against sea star spongivory. Mar. Biol. 143, 443-449. Gagnon, P., Wagner, G., Himmelman, J.H., 2003. Use of a wave tank to study the effects of water motion and algal movement on the displacement of the sea star Asterias vulgaris towards its prey. Mar. Ecol. Prog. Ser. 258, 125-132. Garneau, F.X., Harvey, C, Simard, J.L., Apsimon, J.W., Burnell, D.J., Himmelman, J.H., 1989. The distribution of asterosaponins in various body components ofthe starfish Leptasterias polaris. Comp. Biochem. Phys. B. 92, 411-416. Gamick, E., 1978. Behavioral ecology of Strongylocentrotus droebachiensis (Muller) (Echinodermata: Echinoidea). Aggregating behavior and chemotaxis. Oecologia 37, 77- 84. Gaymer, CF., Himmelman, J.H., 2008. A keystone predatory sea star in the intertidal zone is controlled by a higher-order predatory sea star in the subtidal zone. Mar. Ecol. Prog. Ser. 370,143-153. Gaymer, CF., Himmelman, J.H., Johnson, L.E., 2001a. Distribution and feeding ecology ofthe seastars Leptasterias polaris and Asterias vulgaris in the northern Gulf of St. Lawrence, Canada. J. Mar. Biol. Assoc. U.K. 81, 827-843. Gaymer, CF., Himmelman, J.H., Johnson, L.E., 2001b. Use of prey resources by the seastars Leptasterias polaris and Asterias vulgaris: a comparison between field observations and laboratory experiments. J. Exp. Mar. Biol. Ecol. 262, 13-30. 52

Gaymer, CF., Himmelman, J.H., Johnson, L.E., 2002. Effect of intra- and interspecific interactions on the feeding behavior of two subtidal sea stars. Mar. Ecol. Prog. Ser. 232, 149-162. Godoy, N., Gelcich, S., Vasquez, J.A., Castilla, J.C, 2010. Spearfishing to depletion: evidence from temperate reef fishes in Chile. Ecol. Appl. 20, 1504-1511. Guidetti, P., 2006. Marine reserves reestablish lost predatory interactions and cause community changes in rocky reefs. Ecol. Appl. 16, 963-976. Guidetti, P., Dulcic, J., 2007. Relationships among predatory fish, sea urchins and barrens in Mediterranean rocky reefs across a latitudinal gradient. Mar. Environ. Res. 63, 168-184. Hagen, N.T., Mann, K.H., 1994. Experimental analysis of factors influencing the aggregating behavior ofthe green sea urchin Strongylocentrotus droebachiensis (Muller). J. Exp. Mar. Biol. Ecol. 176, 107-126. Hagen, N.T., Andersen, A., Stabell, O.B., 2002. Alarm responses of the green sea urchin, Strongylocentrotus droebachiensis, induced by chemically labelled durophagous predators and simulated acts of prédation. Mar. Biol. 140, 365-374. Hereu, B., Zabala, M., Linares, C, Sala, E., 2005. The effects of predator abundance and habitat structural complexity on survival of juvenile sea urchins. Mar. Biol. 146, 293-299. Himmelman, J.H., Nedelec, H., 1990. Urchin foraging and algal survival strategies in intensely grazed communities in eastern Canada. Can. J. Fish. Aquat. Sci. 47, 1011-1026. Himmelman, J.H., Dutil, C, 1991. Distribution, population structure and feeding of subtidal seastars in the northern Gulf of St. Lawrence. Mar. Ecol. Prog. Ser. 76, 61-72. Himmelman, J.H., Cardinal, A., Bourget, E., 1983. Community development following removal of urchins, Strongylocentrotus droebachiensis, from the rocky subtidal zone of the St. Lawrence Estuary, Eastern Canada. Oecologia 59, 27-39. Himmelman, J.H., Guderley, H.E., Duncan, P.F., 2009. Responses ofthe saucer scallop Amusium balloti to potential predators. J. Exp. Mar. Biol. Ecol. 378, 58-61. Jackson, G.A., Kiorboe, T., 2004. Zooplankton use of chemodetection to find and eat particles. Mar. Ecol. Prog. Ser. 269, 153-162. Jensen, M., 1966. The response of two sea-urchins to the sea-star Marthasterias glacialis (L.) and other stimuli. Ophelia 3, 209-219. Kats, L.B., Dill, L.M., 1998. The scent of death: chemosensory assessment of prédation risk by prey animals. Ecoscience 5, 361-394. Kerfoot, W., Sih, A., 1987. Prédation: direct and indirect impacts on aquatic communities. University Press of New England, Hanover and London, 389 pp. Krebs, C.J., 2009. Ecology: the experimental analysis of distribution and abundance, San Francisco, 655 pp. Laur, D.R., Ebeling, A.W., Reed, D.C, 1986. Experimental evaluations of substrate types as barriers to sea urchin (Strongylocentrotus spp.) movement. Mar. Biol. 93, 209-215. Legault, C, Himmelman, J.H., 1993. Relation between escape behaviour of benthic marine invertebrates and the risk of prédation. J. Exp. Mar. Biol. Ecol. 170, 55-74. Lima, S.L., Dill, L.M., 1990. Behavioral decisions made under the risk of prédation: a review and prospectus. Can. J. Zool. 68, 619-640. Mahon, A.R., Amsler, CD., McClintock, J.B., Baker, B.J., 2002. Chemo-tactile predator avoidance responses of the common Antarctic limpet Nacella concinna. Polar Biol. 25, 469-473. Mann, K.H., 1982. Kelp, sea urchins and predators: a review of strong interactions in rocky subtidal systems of eastern Canada, 1970-1980. Neth. J. Sea Res. 16, 414-423. 53

Mann, K.H., Wright, J.L.C, Welsford, B.E., Hatfield, E., 1984. Responses of the sea urchin Strongylocentrotus droebachiensis (O.F. Muller) to water-borne stimuli from potential predators and potential food algae. J. Exp. Mar. Biol. Ecol. 79, 233-244. Markowska, M., Kidawa, A., 2007. Encounters between Antarctic limpets, Nacella concinna, and predatory sea stars, Lysasterias sp., in laboratory and field expérimente. Mar. Biol. 151, 1959-1966. Matassa, CM., 2010. Purple sea urchins Strongylocentrotus purpuratus reduce grazing rates in response to risk cues from the spiny lobster Panulirus interruptus. Mar. Ecol. Prog. Ser. 400, 283-288. McClanahan, T.R., Shafir, S.H., 1990. Causes and consequences of sea urchin abundance and diversity in Kenyan coral reef lagoons. Oecologia 83, 362-370. McClintock, J.B., Klinger, T.S., Lawrence, J.M., 1984. Chemoreception in Luidia clathrata (Echinodermata: Asteroidea): qualitative and quantitative aspects of chemotactic responses to low molecular weight compounds. Mar. Biol. 84, 47-52. McClintock, J.B., Angus, R.A., Ho, C.P., Amsler, CD., Baker, B.J., 2008a. Intraspecific agonistic arm-fencing behavior in the Antarctic keystone sea star Odontaster validus influences prey acquisition. Mar. Ecol. Prog. Ser. 371, 297-300. McClintock, J.B., Angus, R.A., Ho, C, Amsler, CD., Baker, B.J., 2008b. A laboratory study of behavioral interactions of the Antarctic keystone sea star Odontaster validus with three sympatrie predatory sea stars. Mar. Biol. 154, 1077-1084. Medina, M., Araya, M., Vega, C, 2004. Feeding and trophies relationships of coastal fish off northern Chile. Invest. Mar. 32, 33-47. Miller, J.A., Byrne, M., 2000. Ceratal autotomy and regeneration in the aeolid nudibranch Phidiana crassicornis and the role of predators. Invertebr. Biol. 119, 167-176. Moitoza, D.J., Phillips, D.W., 1979. Prey defense, predator preference, and nonrandom diet: the interactions between Pycnopodia helianthoides and two species of sea urchins. Mar. Biol. 53, 299-304. Navarrete, S.A., Manzur, T., 2008. Individual- and population-level responses of a keystone predator to geographic variation in prey. Ecology 89, 2005-2018. Nelson, B.V., Vance, R.R., 1979. Diel foraging patterns of the sea urchin Centrostephanus coronatus as a predator avoidance strategy. Mar. Biol. 51, 251-258. Nishizaki, M.T., Ackerman, J.D., 2005. A secondary chemical cue facilitates juvenile-adult postsettlement associations in red sea urchins. Limnol. Oceanogr. 50, 354-362. Paine, R.T., 1980. Food webs: linkage, interaction strength and community infrastructure. J. Anim. Ecol. 49, 667-685. Parker, D.A., Shulman, M.J., 1986. Avoiding prédation: alarm responses of Caribbean sea urchins to simulated prédation on conspecific and heterospecific sea urchins. Mar. Biol. 93, 201-208. Phillips, D.W., 1975. Distance chemoreception-triggered avoidance behavior of the limpets Acmaea (Collisella) limatula and Acmaea (Notoacmea) scutum to the predatory starfish Pisaster ochraceus. J. Exp. Zool. 191, 199-209. Phillips, D.W., 1976. The effect of a species-specific avoidance response to predatory starfish on the intertidal distribution of two gastropods. Oecologia 23, 83-94. Phillips, D.W., 1977. Avoidance and escape responses of gastropod mollusk Olivella biplicata (sowerby) to predatory asteroids. J. Exp. Mar. Biol. Ecol. 28, 77-86. Phillips, D.W., 1978. Chemical mediation of invertebrate defensive behaviors and ability to distinguish between foraging and inactive predators. Mar. Biol. 49, 237-243. 54

Rochette, R., Hamel, J.F., Himmelman, J.H., 1994. Foraging strategy ofthe asteroid Leptasterias polaris: role of prey odors, current and feeding status. Mar. Ecol. Prog. Ser. 106, 93-100. Rochette, R., Arsenault, D.J., Justome, B., Himmelman, J.H., 1998. Chemically-mediated predator-recognition learning in a marine gastropod. Ecoscience 5, 353-360. Rodriguez, S.R., 2003. Consumption of drift kelp by intertidal populations of the sea urchin Tetrapygus niger on the central Chilean coast: possible consequences at different ecological levels. Mar. Ecol. Prog. Ser. 251, 141-151. Rodriguez, S.R., Ojeda, F.P., 1993. Distribution patterns of Tetrapygus niger (Echinodermata: Echinoidea) off the central Chilean coast. Mar. Ecol. Prog. Ser. 101, 157-162. Rodriguez, S.R., Ojeda, F.P., 1998. Behavioral responses ofthe sea urchin Tetrapygus niger to predators and food. Mar. Freshwat. Behav. Physiol. 31, 21-37. Rodriguez, S.R., Farina, J.M., 2001. Effect of drift kelp on the spatial distribution pattern ofthe sea urchin Tetrapygus niger: a geostatistical approach. J. Mar. Biol. Assoc. U.K. 81, 179- 180. Rosenberg, R., Selander, E., 2000. Alarm signal response in the brittle star Amphiura filiformis. Mar. Biol. 136,43-48. Rosenthal, R.J., Chess, J.R., 1972. A predator-prey relationship between the leather star, Dermasterias imbricata, and the purple urchin, Strongylocentrotus purpuratus. Fish B- NOAA 70, 205-216. Ruxton, CD., Sherratt, T.N., Speed, M.P., 2004. Avoiding attack: The evolutionary ecology of crypsis, warning signals and mimicry. Oxford University Press, Oxford. Ryer, C.H., Stoner, A.W., Titgen, R.H., 2004. Behavioral mechanisms underlying the refuge value of benthic habitat structure for two flatfishes with differing anti-predator strategies. Mar. Ecol. Prog. Ser. 268, 231-243. Sala, E., 1997. Fish predators and scavengers ofthe sea urchin Paracentrotus lividus in protected areas ofthe north-west Mediterranean Sea. Mar. Biol. 129, 531-539. Sala, E., Zabala, M., 1996. Fish prédation and the structure of the sea urchin Paracentrotus lividus populations in the NW Mediterranean. Mar. Ecol. Prog. Ser. 140, 71-81. Sala, E., Ribes, M., Hereu, B., Zabala, M., Alva, V., Coma, R., Garrabou, J., 1998. Temporal variability in abundance of the sea urchins Paracentrotus lividus and Arbacia lixula in the northwestern Mediterranean: comparison between a marine reserve and an unprotected area. Mar. Ecol. Prog. Ser. 168, 135-145. Saporito, R.A., Donnelly, M.A., Garraffo, H.M., Spande, T.F., Daly, J.W., 2006. Geographic and seasonal variation in alkaloid-based chemical defenses of Dendrobates pumilio from Bocas del Toro, Panama. J. Chem. Ecol. 32, 795-814. SAS Institute Inc, 2008. SAS/STAT® 9.2 User's Guide. Cary, NC: SAS Institute Inc. Scheibling, R.E., Hamm, J., 1991. Interactions between sea urchins (Strongylocentrotus droebachiensis) and their predators in field and laboratory experiments. Mar. Biol. 110, 105-116. Shears, N.T., Babcock, R.C., 2002. Marine reserves demonstrate top-down control of community structure on temperate reefs. Oecologia 132, 131-142. Sih, A., Wooster, D.E., 1994. Prey behavior, prey dispersal, and predator impacts on stream prey. Ecology 75, 1199-1207. Sih, A., Crowley, P., McPeek, M., Petranka, J., Strohmeier, K, 1985. Prédation, competition, and prey communities: a review of field experiments. Annu. Rev. Ecol. Syst. 16, 269-311. Smith, R.J.F., 1992. Alarm signals in fishes. Rev. Fish Biol. Fish. 2, 33-63. 55

Snedecor, G.W., Cochran, W.G., 1989. Statistical Methods, 8th ed. Iowa State University Press, Iowa. Snyder, N., Snyder, H., 1970. Alarm response oiDiadema antillarum. Science 168, 276-278. Svensen, C, Kiorboe, T., 2000. Remote prey detection in Oithona similis: hydromechanical versus chemical cues. J. Plankton Res. 22, 1155-1166. Tegner, M.J., Levin, L.A., 1983. Spiny lobsters and sea urchins: analysis of a predator-prey interaction. J. Exp. Mar. Biol. Ecol. 73, 125-150. Terlau, H., Shon, K.J., Grilley, M., Stocker, M., Stuhmer, W., Olivera, B.M., 1996. Strategy for rapid immobilization of prey by a fish-hunting marine snail. Nature 381, 148-151. Thompson, M., Drolet, D., Himmelman, J.H., 2005. Localization of infaunal prey by the sea star Leptasterias polaris. Mar. Biol. 146, 887-894. Tokeshi, M., Estrella, C, Paredes, C, 1989. Feeding ecology of a size-structured predator population, the South American sun-star Heliaster helianthus. Mar. Biol. 100, 495-505. Tuya, F., Boyra, A., Sanchez-Jerez, P., Barbera, C, Haroun, R.J., 2004. Relationships between rocky-reef fish assemblages, the sea urchin Diadema antillarum and macroalgae throughout the Canarian Archipelago. Mar. Ecol. Prog. Ser. 278, 157-169. Vadas, R.L., 1977. Preferential feeding: an optimization strategy in sea urchins. Ecol. Monogr. 47,337-371. Vadas, R.L., Elner, R.W., 2003. Responses to prédation cues and food in two species of sympatrie, tropical sea urchins. Mar. Ecol. P.S.Z.N.I. 24, 101-121. Vadas, R.L., Elner, R.W., Garwood, P.E., Babb, I.G., 1986. Experimental evaluation of aggregation behavior in the sea urchin Strongylocentrotus droebachiensis. A reinterpretation. Mar. Biol. 90, 433-448. Vasquez, J.A., Buschmann, A.H., 1997. Herbivore-kelp interactions in Chilean subtidal communities: a review. Rev. Chil. Hist. Nat. 70, 41-52. Vega, J.M.A., Vasquez, J.A., Buschmann, A.H., 2005. Population biology of the subtidal Macrocystis integrifolia and Lessonia trabeculata (Laminariales, Phaeophyceae) in an upwelling ecosystem of northern Chile: interannual variability and El Nino 1997-1998. Rev. Chil. Hist. Nat. 78, 33-50. Viviani, C.A., 1978. Prédation interespecifica, canibalismo y autotomia como mecanismo de escape en las especies de Asteroidea (Echinodermata) en el litoral del desierto del norte grande de Chile. Universidad del Norte, Iquique. Weightman, J.O., Arsenault, D.J., 2002. Predator classification by the sea pen Ptilosarcus gurneyi (Cnidaria): role of waterborne chemical cues and physical contact with predatory sea stars. Can. J. Zool. 80, 185-190.