Dynamique de répartition et mécanismes de défense des ... · III Abstract In the northern Gulf of St. Lawrence, the green sea urchin, Strongylocentrotus droebachiensis, is the

Patrick Gagnon

Dynamique de répartition et mécanismes de défense des macrophytes de la zone infralittorale rocheuse dominée par

l’oursin vert, Strongylocentrotus droebachiensis

Thèse présentée

à la Faculté des études supérieures de lʹUniversité Laval pour l’obtention

du grade de Philosophiae Doctor (Ph.D.)

Département de biologie FACULTÉ DES SCIENCES ET DE GÉNIE

UNIVERSITÉ LAVAL QUÉBEC

DÉCEMBRE 2003

© Patrick Gagnon, 2003

À mes très chers parents, Patricia Bourget et Richard Gagnon,

ainsi qu’à Nellie…

I

Sommaire

Le but de mon étude était de caractériser la dynamique de répartition des

macrophytes de la zone infralittorale rocheuse de l’archipel de Mingan, au nord du

golfe du Saint‐Laurent, puis d’examiner les mécanismes de défense permettant à

certains macrophytes dominants de résister au broutage de l’oursin vert,

Strongylocentrotus droebachiensis. La position de la limite inférieure des ceintures de

laminaires variait grandement dans le temps et d’un site à l’autre et était

principalement contrôlée par le broutage de l’oursin. En zone dénudée, une grande

stabilité temporelle caractérisait la distribution et l’abondance de la phaéophycée

pérenne Agarum cribrosum dont l’agrégation en plaques résilientes à de faibles

perturbations favorisait le recrutement algal. L’évitement du broutage de l’oursin par

les juvéniles d’A. cribrosum reposait sur leur association avec la phaéophycée annuelle

Desmarestia viridis. La répulsion de l’oursin par D. viridis était principalement

attribuable au mouvement de balayage de l’algue, résultat de l’effet combiné de sa

délicate morphologie et de l’action des vagues.

Auteur: Patrick Gagnon

Directeur: John H. Himmelman

II

Résumé

Dans le nord du golfe du Saint‐Laurent, lʹoursin vert, Strongylocentrotus droebachiensis,

est lʹherbivore dominant en zone infralittorale rocheuse. Son impact sur la structure et

la dynamique des communautés de macrophytes est toutefois peu connu. Le but de

mon étude était de caractériser la dynamique de répartition des macrophytes de la

zone infralittorale rocheuse de l’archipel de Mingan, puis d’examiner les mécanismes

de défense permettant à certains macrophytes dominants de résister au broutage de

l’oursin. Des observations et expériences sur huit sites sur des périodes couvrant

jusqu’à deux ans ont indiqué que la position de la limite inférieure des ceintures de

laminaires variait grandement dans le temps et entre les sites et était principalement

contrôlée par le broutage de l’oursin. Lʹabrasion par les glaces et la formation de

barrières algales limitant le déplacement des oursins ont vraisemblablement affecté

cette variation. En zone dénudée, une grande stabilité temporelle caractérisait la

distribution et l’abondance de la phaéophycée pérenne Agarum cribrosum.

L’agrégation d’A. cribrosum en plaques résilientes à de faibles perturbations favorisait

le recrutement algal, dont la rhodophycée Ptilota serrata. L’évitement du broutage de

l’oursin par les juvéniles d’A. cribrosum reposait sur leur association avec la

phaéophycée annuelle Desmarestia viridis. Pour les juvéniles évoluant sous un couvert

de D. viridis, le coût d’une perte de croissance était nettement compensé par le

bénéfice d’une réduction du broutage. Des expériences en laboratoire ont établi que la

répulsion de l’oursin par D. viridis était principalement due au mouvement de

balayage de l’algue, résultat de l’effet combiné de sa délicate morphologie et de

l’action des vagues puis, dans une faible mesure, de sa composition chimique. Les

attaques de l’oursin sur D. viridis augmentaient avec la densité de l’oursin mais

diminuaient avec l’augmentation du mouvement de l’eau.



III

Abstract

In the northern Gulf of St. Lawrence, the green sea urchin, Strongylocentrotus

droebachiensis, is the dominant herbivore in the rocky subtidal zone. Its impact on the

structure and dynamics of algal communities is, however, poorly known. The

objective of my study was to characterize the dynamics in the distribution and

abundance of macrophytes in the rocky subtidal zone in the Mingan Islands, and to

examine the mechanisms of defense that allow dominant macrophytes to resist urchin

grazing. Observations and experiments at eight sites over periods of up to two years

indicated that the position of the lower limit of the kelp beds varied markedly

through time and among sites and was largely controlled by urchin grazing. Ice

souring and the formation of algal barriers that reduced urchin displacement likely

affected this variation. In the barrens zone, there was a high temporal stability in the

distribution and abundance of the perennial phaeophyte Agarum cribrosum. The

aggregation of A. cribrosum into patches resilient to small disturbances enhanced algal

recruitment, notably that of the rhodophyte Ptilota serrata. The avoidance of urchin

grazing by juveniles of A. cribrosum was related to their association with the annual

phaeophyte Desmarestia viridis. For juveniles, the cost of a reduced growth under a

D. viridis canopy was clearly outweighed by the benefit of reduced urchin grazing.

Laboratory experiments established that the repulsion of urchins by D. viridis was

primarily due to the sweeping motion of the alga resulting from the interaction

between its delicately‐branched morphology and wave action and, to a lesser extent,

by its chemical make‐up. Urchin attacks on D. viridis increased with urchin density,

but decreased with increased water motion.



IV

Avant‐propos

Cette thèse comporte six chapitres. Les chapitres I et VI correspondent

respectivement aux introduction et conclusion générales. Les chapitres II, III, IV et V

(corps de la thèse) font état de mes travaux de recherche dans l’archipel de Mingan et

ont été rédigés en anglais, sous forme d’articles scientifiques dont voici les références:

Chapitre II: Gagnon P., Himmelman J.H. & Johnson L.E. (sous presse). Temporal

variation in community interfaces: kelp bed boundary dynamics

adjacent to persistent urchin barrens. Mar. Biol.

Chapitre III: Gagnon P., Himmelman J.H. & Johnson L.E. Spatial and temporal

stability in algal assemblages on urchin barrens in the northern Gulf of

St. Lawrence.

Chapitre IV: Gagnon P., Himmelman J.H. & Johnson L.E. (2003). Algal colonization

in urchin barrens: defense by association during recruitment of the

brown alga Agarum cribrosum. J. Exp. Mar. Biol. Ecol. 290: 179‐196.

Chapitre V: Gagnon P., St‐Hilaire‐Gravel L.V., Himmelman J.H. & Johnson L.E.

Multiple plant defenses against herbivores: the interplay of chemical,

morphological, and hydrodynamic factors.

J’ai conçu et réalisé toutes les expériences et observations de terrain et de

laboratoire, effectué la totalité des analyses statistiques (avec l’appui de statisticiens

pour les analyses plus complexes), puis rédigé les différents chapitres avec l’aide de

mes codirecteurs John Himmelman et Ladd Johnson. Les résultats apparaissant dans

cette thèse ont été présentés lors des congrès et ateliers suivants:

Gagnon P., Himmelman J.H. & Johnson L.E. (2003). Sweeping or using acid: how

does Desmarestia viridis deter urchin attacks? 32nd Benthic Ecology Meeting,

Groton, Connecticut, USA.

V

Gagnon P., Himmelman J.H. & Johnson L.E. (2000). Seaweed survival in barren

zones: a case of associational refuge for juvenile Agarum cribrosum.

29th Benthic Ecology Meeting, Wilmington, North Carolina, USA.

Gagnon P., Himmelman J.H. & Johnson L.E. (2000). Utilisation du couvert de

Desmarestia viridis par les juvéniles d’Agarum cribrosum pour une protection

contre l’herbivorie. 68ème Congrès International de l’ACFAS (Association

Canadienne Française pour l’Avancement des Sciences), Montréal, Québec,

Canada.

Gagnon P., Himmelman J.H. & Johnson L.E. (2000). Utilisation du couvert de

Desmarestia viridis par les juvéniles d’Agarum cribrosum pour une protection

contre l’herbivorie. 25ème Congrès annuel de la Société Québécoise pour

l’Étude Biologique du Comportement, Rimouski, Québec, Canada.

Gagnon, P. (1999). Importance du refuge d’association pour l’établissement des

macrophytes de la communauté infralittorale. Forum Québécois en Sciences

de la Mer, Mont‐Joli, Québec, Canada.

La description du générateur de vagues que j’ai conçu et fabriqué pour réaliser

certaines de mes expériences apparaît en annexe du chapitre V, puis dans l’article

suivant rédigé simultanément à mes travaux de thèse:

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.

Je remercie mon directeur, John Himmelman, ainsi que mon codirecteur,

Ladd Johnson, pour la confiance témoignée et la latitude accordée depuis mes tous

premiers pas à l’initiation à la recherche, jusqu’à la fin de cette aventure doctorale.

Leur disponibilité, leur support et leurs nombreuses recommandations scientifiques

pendant la rédaction de ma thèse ont été particulièrement appréciés et la rigueur de

leur encadrement très formatrice. Merci également à Gilles Houle, Jeremy McNeil et

VI

Edwin Bourget pour leur grande ouverture et pour toutes ces discussions profitables,

puis au Conseil de Recherche en Sciences Naturelles et Génie du Canada (CRSNG),

au Fonds pour la formation de Chercheurs et l’Aide à la Recherche du Québec

(FCAR), au Ministère des Pêches et Océans du Canada (MPO), à Québec‐Océan

(GIROQ) et au Département de biologie de l’Université Laval pour leur généreuse

contribution monétaire.

J’exprime toute ma gratitude aux personnes qui ont participé, de près ou de

loin, à la réalisation de mon étude. Merci à mes précieux assistants de recherche pour

toutes ces heures passées au laboratoire et pour ces interminables plongées dans des

eaux glaciales et agitées: Philippe Gauthier, Valérie Messier, Brigitte Laberge,

François Durocher, Simon‐Pierre Gingras, David Drolet, Damien Aiello,

Miles Thompson, Erasto Marcano, Isabelle Deschênes, Marc‐Olivier Nadon et

Catherine Vallières. Un merci tout particulier à Oliver D’Amours, Giselle Wagner,

Louis Vincent St‐Hilaire‐Gravel et Mark Dionne pour leur énergie sans borne, leur

enthousiasme contagieux et la qualité du travail accompli, à mes collègues de

laboratoire Jean‐François Raymond, Julien Gaudette, Carlos Gaymer, Martin Guay et

Thierry Gosselin pour les étés de terrains exceptionnels passés en leur compagnie à

Havre‐Saint‐Pierre, puis à Benoît Roberge et Nancy Dénommée de Parcs Canada

pour le temps consacré à partager avec moi le fruit de leurs observations sur la

formation hivernale des glaces dans l’archipel de Mingan.

Un gros merci à Jean‐Marc Daigle, ainsi qu’à son frère Gaétan, puis à

Eva Bessac pour leur exceptionnelle patience à m’aider dans mes analyses et à me

transmettre, pendant d’innombrables heures, une partie de leur grande connaissance

de ce monde fascinant que constituent les statistiques.

Jamais je n’exprimerai assez de reconnaissance envers ma mère Patricia, mon

père Richard, mon frère Gabriel, puis ma sœur Sophie qui ont été de tous les instants

présents. Votre sensibilité, vos encouragements, votre support et votre aide à tous les

niveaux ont nettement dépassé tout ce que je ne pouvais même oser espérer. Pour

leur accueil si chaleureux à Havre‐Saint‐Pierre, je remercie grandement Michelle

VII

Maloney et Clément Cormier dont la gentillesse n’a d’égal que la grandeur de leur

cœur. Merci aussi à mes bons amis Louis‐Philippe, Geneviève, Gaétan, Olivier, Paul,

Jean‐Claude, Roger, Stéphane, Mélanie, Hugo, Catherine et Annie‐Claude pour leurs

encouragements soutenus et leur compréhension. Vous étiez bien présents à mes

côtés malgré la distance…

Mais par‐dessus tout, j’offre mes plus chaleureux remerciements à ma très

chère Nellie, l’étoile de ma vie et ma confidente, pour avoir brillamment dirigé la

barque lors des moments de brouillard et de grandes vagues et pour son

extraordinaire patience, sa douceur, son entrain, sa compréhension et son support

moral et affectif inconditionnels… une femme tout à fait exceptionnelle que j’admire

profondément.

Merci à toutes et à tous!

VIII

Table des matières

Sommaire ..............................................................................................................................I

Résumé ............................................................................................................................ II

Abstract .......................................................................................................................... III

Avant‐propos .......................................................................................................................... IV

Table des matières ................................................................................................................. VIII

Chapitre I. Introduction générale...................................................................................... 1

Chapitre II Temporal variation in community interfaces: kelp bed boundary dynamics adjacent to persistent urchin barrens ...................................... 11

Résumé ..................................................................................................................................... 12 Abstract .................................................................................................................................... 13 Introduction ............................................................................................................................. 14 Materials and methods........................................................................................................... 18

Study sites ...................................................................................................................... 18 Temporal variability in A. esculenta beds and urchin populations........................ 18 Effect of urchins on algal recruitment in the barrens .............................................. 21 Limitation of upward movements of urchins by A. cribrosum ............................... 22 Statistical analysis ......................................................................................................... 23

Results....................................................................................................................................... 24 Temporal variability in the lower limit of kelp beds............................................... 24 Structure of kelp beds................................................................................................... 27 Structure of urchin populations.................................................................................. 30 Effect of urchins on algal recruitment in the barrens .............................................. 34 Limitation of upward movements of urchins by A. cribrosum ............................... 36

Discussion ................................................................................................................................ 36

IX

Chapitre III. Spatial and temporal stability in algal assemblages on urchin

barrens in the northern Gulf of St. Lawrence .......................................... 44


Temporal variability in abundance and distribution of A. cribrosum.................... 49 Temporal variability in the size of A. cribrosum stands........................................... 50 Distribution of algae and urchins near A. cribrosum ................................................ 50 Effect of A. cribrosum canopy on recruitment and abundance of algae ................ 51 Statistical analysis ......................................................................................................... 51

Results....................................................................................................................................... 53 Temporal variability in abundance and distribution of A. cribrosum.................... 53 Temporal variability in the size of A. cribrosum stands........................................... 56 Distribution of algae and urchins near A. cribrosum ................................................ 56 Effect of A. cribrosum canopy on recruitment and abundance of algae ................ 56

Discussion ................................................................................................................................ 63

Chapitre IV. Algal colonization in urchin barrens: defense by association during recruitment of the brown alga Agarum cribrosum ..................... 67


Study sites ...................................................................................................................... 72 Distribution of juvenile A. cribrosum and urchins near D. viridis........................... 73 Protection provided by D. viridis ................................................................................ 74 Growth of juvenile A. cribrosum.................................................................................. 75 Seasonal cycle of D. viridis ........................................................................................... 76 Statistical analysis ......................................................................................................... 76

Results....................................................................................................................................... 78 Distribution of juvenile A. cribrosum and urchins near D. viridis........................... 78 Protection provided by D. viridis ................................................................................ 81 Growth of juvenile A. cribrosum.................................................................................. 85 Seasonal cycle of D. viridis ........................................................................................... 85

Discussion ................................................................................................................................ 88

X

Chapitre V. Multiple plant defenses against herbivores: the interplay of

chemical, morphological, and hydrodynamic factors............................. 93


Response of urchins to odors of Desmarestia viridis ............................................... 100 Mechanical repulsion of urchins by Desmarestia viridis......................................... 101 Effect of urchin density on attack success on Desmarestia viridis ......................... 103 Field observations ....................................................................................................... 103 Statistical analysis ....................................................................................................... 103

Results..................................................................................................................................... 105 Response of urchins to odors of Desmarestia viridis ............................................... 105 Mechanical repulsion of urchins by Desmarestia viridis......................................... 108 Effect of urchin density on attack success on Desmarestia viridis ......................... 111 Field observations ....................................................................................................... 111

Discussion .............................................................................................................................. 111 Appendix 5.1.......................................................................................................................... 118

Chapitre VI. Conclusion générale ................................................................................... 119

Dynamique de répartition des macrophytes .......................................................... 120 Mécanismes de défense des macrophytes............................................................... 121 Portée de l’étude ......................................................................................................... 122 Nouvelles questions et hypothèses .......................................................................... 124

Bibliographie ......................................................................................................................... 127

Chapitre I

Introduction générale

2

Toute communauté écologique est fondamentalement dynamique (hétérogène dans

le temps et l’espace) car les organismes qui la composent sont en constante interaction

(entre eux et avec les propriétés changeantes de leur environnement physico‐

chimique) et s’affectent dans leur distribution et leur abondance relatives (Connell,

1961; Paine, 1966; Holling, 1973; Schoener, 1974; Branch, 1975; Menge, 1976; Sousa,

1984; Sih et al., 1985). La stabilité d’une communauté réfère à sa capacité de

conservation de son état (composition) malgré la force exercée par une perturbation

(Holling, 1973; Krebs, 2001). Les perturbations (aléatoires ou cycliques) sous forme de

variations dans les conditions physiques (e.g., feu, vent, vagues), d’invasions

d’espèces compétitrices et d’ennemis naturels peuvent déplacer l’état d’équilibre

(avant perturbation) d’une communauté vers un état stable alternatif

(Connell & Slatyer, 1977; White, 1979; Ebeling et al., 1985; Chapman & Johnson, 1990;

Vadas et al., 1990; Dayton et al., 1992; Trowbridge, 1995; Ruiz et al., 1997;

Underwood, 1999). Cet état d’équilibre (climax) est l’aboutissement d’un patron de

colonisation et d’extinction des espèces (succession) non saisonnier, directionnel et

continu (Odum, 1969; Begon et al., 1996) dirigé par un complexe de processus

interactifs (e.g., prédation, compétition, immigration et émigration) parmi lesquels

certains en neutralisent d’autres (Odum, 1969; Underwood & Jernakoff, 1981;

Witman, 1987; Hendrix et al., 1988; Davidson, 1993; Benedetti‐Cecchi, 2000;

Duffy & Hay, 2000). Une communauté n’est donc jamais à un point d’équilibre exact

car l’état de celle‐ci se retrouve dans un flux constant (Lewontin, 1969; Odum, 1969;

Holling, 1973; Sutherland, 1974; Gray, 1977).

L’herbivorie peut exercer un impact considérable sur la structure et la

dynamique des communautés végétales (Ogden & Lobel, 1978; Witman, 1987;

Lindroth, 1989; Weis & Berenbaum, 1989; Brawley, 1992; Vadas & Elner, 1992;

Vasquez & Buschmann, 1997; Zamora et al., 1999; Duffy & Hay, 2001) tel que

démontré par l’effet dramatique d’herbivores sur la distribution et l’abondance des

organismes végétaux (Jones & Kain, 1967; Breen & Mann, 1976b; Dean et al., 1984;

Carpenter, 1986; Lewis, 1986; Morrison, 1988; Lindroth, 1989; Duffy & Hay, 2000).

Dans une communauté où le recrutement et la survie des végétaux sont limités par

3

l’herbivorie, la réduction de l’abondance des herbivores par une perturbation

(e.g., tempêtes, maladies et prédation) peut provoquer le déclenchement d’une

succession végétale secondaire menant à l’établissement d’une communauté dont la

richesse et la biomasse spécifiques diffèrent grandement de celles de la communauté

maintenue par les herbivores (Paine & Vadas, 1969; Pearse & Hines, 1979; Miller,

1985; Lindroth, 1989; Brown & Gange, 1992; Sousa & Connell, 1992;

Leinaas & Christie, 1996; Scheibling et al., 1999). L’impact d’un herbivore sur une

communauté végétale est non seulement fonction de son abondance, mais également

de sa taille par rapport aux organismes végétaux, de ses interactions avec les

compétiteurs et les prédateurs, ainsi que de l’abondance, de la productivité

(recrutement ou croissance) et de la richesse spécifique des organismes végétaux

(Brown & Ewel, 1987; Sousa & Connell, 1992). L’état d’une communauté végétale

repose donc en partie sur un complexe d’interactions plantes‐herbivores à l’intérieur

desquelles l’exploitation des ressources végétales par un herbivore particulier est

déterminée par les conditions environnementales (e.g., température, vents, courants),

la disponibilité de la ressource, puis les attributs individuels des plantes

(e.g., digestibilité et valeur énergétique) et de l’herbivore (e.g., degré de mobilité,

préférences alimentaires) (Emlen, 1966; MacArthur & Pianka, 1966; Menge, 1972;

Charnov, 1976).

On distingue par ailleurs deux composantes dans le comportement de quête

alimentaire d’un animal (herbivore): (1) le comportement qui lui permet d’entrer en

contact avec sa proie (une plante) et (2) sa réponse à cette source de nourriture une

fois le contact établi (Chapman & Underwood, 1992). Comme les plantes ne peuvent

se déplacer et fuir en cas de besoin, leur vulnérabilité aux attaques d’un herbivore est

particulièrement élevée et le contact entre l’herbivore et la plante est donc facilité par

rapport à une proie mobile. À ce titre, les herbivores sont d’importants agents de

sélection contre lesquels les plantes ont développé des adaptations structurales et

métaboliques remarquables qui diminuent cette vulnérabilité en modulant la réponse

alimentaire de l’herbivore (Lindroth, 1989; Weis & Berenbaum, 1989; Hay & Fenical,

1992; Coley & Barone, 1996; Duffy & Hay, 2001). Les plantes utilisent l’un ou l’autre

4

de deux mécanismes de résistance à l’herbivorie: la tolérance et l’évitement

(Duffy & Hay, 1990; Rosenthal & Kotanen, 1994). Une plante peut tolérer la présence

d’un herbivore grâce à un taux de croissance supérieur (surcompensatoire) au taux de

broutage de l’herbivore, à une activité photosynthétique accrue, à l’utilisation de

produits de réserves et au retard de certaines étapes de sa phénologie (McNaughton,

1983; Lindroth, 1989; Davidson, 1993; Tiffin, 2000). Pour éviter un herbivore, une

plante peut (1) produire des défenses structurales (e.g., aiguilles, épines, tissus

calcifiés, Steneck, 1985; Myers & Bazely, 1991) et chimiques (e.g., métabolites

secondaires comprenant terpènes, alkaloïdes et phénols, Hay & Fenical, 1988;

Karban & Myers, 1989; Schappert & Shore, 1999; Cetrulo & Hay, 2000; Pavia & Toth,

2000; Becerra et al., 2001; Schnitzler et al., 2001) permettant de repousser l’herbivore

ou de réduire son taux d’alimentation ou (2) utiliser un refuge temporel

et/ou bénéficier d’un refuge spatial réduisant les probabilités de contact avec

l’herbivore (Weis & Berenbaum, 1989).

Un cas particulier de refuge spatio‐temporel, connu sous le nom de refuge

(ou défense) par association, survient lorsqu’une plante échappe à l’herbivorie en

s’associant à d’autres plantes (ou animaux) plus ou moins attirantes pour l’herbivore

(Tahvanainen & Root, 1972; Littler et al., 1987; Pfister & Hay, 1988; Duffy & Hay,

1990; Kerr & Paul, 1995). L’association peut accroître la compétition pour les

ressources (e.g., espace, lumière, nutriments) et engendrer une perte de croissance

pour l’une ou l’autre des espèces végétales impliquées (Hay, 1986; McAuliffe, 1986;

Pfister & Hay, 1988; Forseth et al., 2001). Pour l’espèce protégée, cette perte est

cependant compensée par le gain de protection permettant la survie. La majorité des

études sur les mécanismes de défense des plantes impliquent des individus adultes.

Or, les juvéniles sont particulièrement vulnérables aux attaques des herbivores car, en

général, leurs attributs structurels et chimiques répulsifs ne sont pas assez développés

pour procurer une défense efficace contre les herbivores (Davidson, 1993;

Van Alstyne et al., 1999) et ils ne disposent pas d’une taille refuge (sensu Paine, 1976;

Lubchenco & Gaines, 1981) pouvant gêner le broutage ou minimiser l’impact de

5

pertes tissulaires. La défense par association peut donc s’avérer déterminante pour la

survie des juvéniles et l’évitement de l’extinction locale d’une espèce.

L’oursin vert, Strongylocentrotus droebachiensis, peuple les fonds des régions

arctiques‐boréales des océans Atlantique et Pacifique (Mortensen, 1943; Jensen, 1974;

Bazhin, 1998). Bien qu’on le retrouve de la limite des marées basses jusqu’à 300 m de

profondeur, il est plus commun en zone peu profonde, à moins de 50 m de

profondeur (Jensen, 1974; Scheibling & Hatcher, 2001) et généralement associé à des

substrats de type rocheux (e.g., roche‐mère, blocs, cailloux, Himmelman, 1986;

Scheibling & Raymond, 1990). L’oursin vert est principalement herbivore mais

consomme également des carcasses animales ainsi que d’autres invertébrés et des

proies bactériennes (Himmelman & Steele, 1971; Lawrence, 1975; Duggins, 1981;

Johnson & Mann, 1982; Briscoe & Sebens, 1988; Nestler & Harris, 1994). Il tolère, sans

mal apparent, des périodes de jeûne prolongé (jusqu’à un mois, Garnick, 1978). Il

utilise la chémodétection à distance pour détecter sa nourriture et affiche un haut

degré de sélectivité (préférence) alimentaire (Vadas, 1977; Garnick, 1978; Larson et al.,

1980; Sloan & Campbell, 1982; Mann et al., 1984; Himmelman & Nédélec, 1990; Prince

& LeBlanc, 1992). Cette sélectivité est cependant fortement influencée par des facteurs

environnementaux tels que la disponibilité des ressources, l’action du courant et des

vagues, la température, la compétition et la prédation (Lawrence, 1975; Vadas, 1977;

Himmelman, 1984; Vadas, 1990; Scheibling, 1996). Le taux d’alimentation individuel

de S. droebachiensis fluctue de façon saisonnière, habituellement avec un pic en

période estivale et un creux en période hivernale (Vadas, 1977; Larson et al., 1980;

Himmelman, 1984; Meidel & Scheibling, 1998). La production et la maturation des

tissus gonadiques débutent généralement à l’automne et l’expulsion des gamètes

survient en début de printemps, vraisemblablement en conjonction avec

l’augmentation rapide de la production phytoplanctonique (Himmelman, 1975;

Falk‐Petersen & Lonning, 1983; Starr et al., 1990; Meidel & Scheibling, 1998).

Dans le nord‐ouest de l’Atlantique, S. droebachiensis exerce une forte influence

sur la distribution et l’abondance des macrophytes (Vadas & Elner, 1992). Le

6

développement rapide (quelques mois) d’assemblages végétaux luxuriants à la suite

d’exclusions manuelles et naturelles de l’oursin témoigne de son impact négatif sur le

recrutement et la survie algale (Himmelman et al., 1983; Scheibling, 1986; Keats et al.,

1990). Néanmoins, les portions côtières rocheuses peu profondes exposées aux

vagues supportent habituellement de vastes et compactes bandes végétales dominées

par les laminaires, communément appelées ceintures de laminaires (Vadas & Elner,

1992; Lobban & Harrison, 1994; Graham & Wilcox, 2000; Tegner & Dayton, 2000).

L’importance écologique de ces ceintures est énorme car en plus de figurer parmi les

systèmes les plus productifs au monde (Mann, 1973), elles procurent un habitat à

plusieurs espèces pour le recrutement, l’alimentation et la reproduction, de même

qu’un refuge contre les prédateurs et l’action des vagues et des courants

(Himmelman et al., 1983; Himmelman & Lavergne, 1985; Vadas & Elner, 1992;

Bologna & Steneck, 1993; Vasquez, 1993; Bégin, 2002; Gagnon et al., 2003a). Des

fluctuations substantielles au sein des populations d’oursins pourraient donc avoir de

profondes répercussions (positives ou négatives) sur les communautés végétales

(e.g., ceintures de laminaires) et, en bout de ligne, sur la structure du reste de la

communauté infralittorale. Tel est le cas sur la côte atlantique de la Nouvelle‐Écosse:

la recrudescence épisodique d’un pathogène (amibe) provoque l’effondrement massif

des populations d’oursins et résulte en l’expansion rapide des ceintures de laminaires

(Miller & Colodey, 1983; Scheibling & Stephenson, 1984; Scheibling, 1986;

Scheibling & Hennigar, 1997).

Dans l’archipel de Mingan, au nord du golfe du Saint‐Laurent, S. droebachiensis

domine la zone infralittorale rocheuse peu profonde où sa densité est, en général, très

élevée (plus de 300 individus m‐2, Himmelman & Nédélec, 1990). Cette zone se

caractérise par un paysage subdivisé en deux communautés végétales qui diffèrent

nettement en terme de biomasse et de richesse spécifiques. L’une d’elles, la zone des

laminaires, occupe les premiers mètres d’eau et est composée d’un dense assemblage

végétal dominé par Alaria esculenta et, de façon moins importante, par d’autres

espèces du groupe des laminaires (e.g., Laminaria spp., Sacchoriza dermatodea).

À l’inverse, la zone dénudée, attenante à la zone des laminaires, est composée d’une

7

mosaïque de plaques végétales parsemant de vastes tapis d’algues calcaires

(e.g., Lithothamnion sp., Clathromorphum sp.) résistantes au broutage (Steneck, 1983;

Steneck, 1985; Steneck, 1986). L’abondance de l’oursin y est nettement plus élevée que

dans la zone des laminaires. La transition entre les deux communautés végétales est

habituellement abrupte et coïncide avec la présence de fronts d’oursins compacts à la

limite inférieure des ceintures de laminaires.

Les phaéophycées Agarum cribrosum et Desmarestia viridis sont les seules

macroalgues charnues communément rencontrées en zone dénudée, ce qui soulève

d’intéressantes questions quant aux attributs et mécanismes leur permettant de

croître dans cet environnement que l’herbivorie par l’oursin rend si hostile à la

végétation. La laminaire pérenne, A. cribrosum, possède un large crampon, un stipe

semi‐rigide et une large fronde ondulée. On la retrouve principalement en petits

agrégats (1‐10 m2), bien qu’elle forme à l’occasion de larges bandes (>500 m2),

particulièrement en bordure de la limite inférieure des ceintures de laminaires.

Sa croissance est lente comparativement aux autres espèces de laminaires

(Vadas, 1968). La présence de composés chimiques répulsifs dans les tissus

d’A. cribrosum (e.g, phénols) expliquerait sa faible consommation par l’oursin vert

(Vadas, 1977; Larson et al., 1980; Himmelman & Nédélec, 1990). Si les adultes

d’A. cribrosum apparaissent peu vulnérables à la prédation par l’oursin, des

observations exploratoires de la distribution des juvéniles suggèrent que ces derniers

le sont grandement. Peu de juvéniles sont localisés aux endroits facilement accessibles

par l’oursin (e.g., aires ouvertes). Ils semblent pour la plupart confinés à l’intérieur

d’agrégats de conspécifiques adultes et sous le couvert de D. viridis. Le recrutement et

la survie des juvéniles d’A. cribrosum pourrait donc dépendre d’associations algales

intra‐ et interspécifiques contre le broutage de l’oursin (e.g., refuge par association).

À l’inverse, D. viridis est une espèce annuelle dotée d’un petit crampon, d’un

stipe hautement flexible et d’une large fronde extrêmement ramifiée. La production et

l’accumulation par D. viridis d’acide sulfurique à l’intérieur de vacuoles spécialisées

la rend très acide (pH 0.5‐0.9, McClintock et al., 1982; Sasaki et al., 1999). L’utilité

8

défensive d’une telle acidité contre l’herbivorie (défense chimique) a été suggérée à

maintes reprises chez plusieurs espèces de Desmarestiales (dont D. viridis), sans

toutefois avoir fait l’objet d’investigations rigoureuses (Anderson & Velimirov, 1982;

Dayton, 1985a; Thompson, 1988; Himmelman & Nédélec, 1990; Lobban & Harrison,

1994). Par ailleurs, des observations in situ ont indiqué que l’oursin semble éviter le

contact avec des frondes de D. viridis agitées par les vagues mais qu’il s’agrège

rapidement, en eau calme, sur des frondes immobiles (Himmelman & Steele, 1971;

Himmelman, 1984; Konar, 2000). Le mouvement de balayage de l’algue causé par

l’action des vagues pourrait donc également contribuer à maintenir l’oursin à

distance (défense mécanique). Une investigation de la réponse comportementale de

S. droebachiensis face à D. viridis permettrait de déterminer la contribution d’une

défense chimique et/ou mécanique à la capacité répulsive de l’algue contre l’oursin.

Aucune étude n’a encore quantifié l’impact de l’oursin vert sur la structure et

la dynamique de répartition naturelle des assemblages algaux dans le nord du golfe

du Saint‐Laurent. Certes, les quelques études réalisées dans (ou près de) cette région

appuient le concept de sélectivité alimentaire chez l’oursin et soulignent son effet

négatif sur le recrutement et la survie algale (Himmelman, 1969; Himmelman et al.,

1983; Himmelman, 1984; Himmelman & Nédélec, 1990; Keats et al., 1990). Cependant,

(1) la nature ponctuelle de ces études, (2) l’étroitesse des échelles spatiales impliquées,

(3) les approches expérimentales utilisées (perturbations anthropiques) et (4) le

manque de relativisation de l’impact de l’oursin sur la structure et la dynamique des

communautés de macrophytes par rapport à l’impact d’autres facteurs (e.g., abrasion

par les glaces, force des courants, compétition intra‐ et interspécifique) ne permettent

pas de dresser de portrait rigoureux des mécanismes écologiques qui alimentent la

dynamique communautaire naturelle et, encore moins, de quantifier cette dernière.

Plusieurs interrogations fondamentales ont motivé la tenue de la présente

étude sur les interactions algues‐oursins dans le nord du golfe du Saint‐Laurent.

Notamment, quels patrons de variation temporelle caractérisent la structure et la

distribution des communautés de macrophytes et des populations d’oursins?

9

Ces patrons s’appliquent‐ils uniformément à l’ensemble des localités ou varient‐ils

plutôt d’un endroit à l’autre? Dans quelle mesure le broutage de l’oursin affecte‐il la

stabilité des communautés végétales et quelle est son importance par rapport à

d’autres facteurs? Finalement, quels sont les mécanismes qui permettent à certains

macrophytes de résister au broutage de l’oursin? En l’absence de réponses à ces

questions, notre connaissance des facteurs et processus écologiques qui façonnent cet

écosystème est limitée, d’où la nécessité d’entreprendre l’examen attentif des

interactions qu’entretiennent l’oursin vert et les communautés de producteurs

primaires.

L’objectif de mon étude est double: (1) caractériser la dynamique de répartition

des macrophytes de la zone infralittorale et identifier les principaux facteurs à

l’origine de cette dynamique dans l’archipel de Mingan, puis (2) examiner les

mécanismes défensifs utilisés par certains macrophytes dominants contre le broutage

de l’oursin vert. Deux chapitres caractérisent la dynamique de répartition des

assemblages algaux, l’un au voisinage de l’interface entre la zone des laminaires et la

zone dénudée, l’autre dans la zone dénudée, puis deux autres chapitres examinent les

interactions entre l’oursin vert et les phaéophycées D. viridis et A. cribroum.

Mes travaux résultent d’une combinaison d’observations et d’expériences effectuées

en milieu naturel en utilisant le scaphandre autonome (SCUBA), puis en laboratoire, à

l’aide de dispositifs et d’appareils décrits dans la littérature scientifique

(e.g., labyrinthe en Y) ou que j’ai conçus afin de contrôler adéquatement les

paramètres de certaines expériences qui ne pouvaient être réalisées en milieu naturel

(e.g., générateur de vagues oscillatoires). Le chapitre II examine les facteurs affectant

la stabilité temporelle de la limite inférieure des ceintures de laminaires en accordant

une attention toute particulière aux impacts du broutage de l’oursin vert. Le

chapitre III évalue la stabilité temporelle de la distribution et de l’abondance des

assemblages algaux de la zone dénudée dominés par A. cribrosum. Le chapitre IV teste

l’hypothèse d’une défense par association permettant aux juvéniles d’A. cribrosum

d’éviter le broutage de l’oursin vert. Finalement, le chapitre V examine les

10

mécanismes par lesquels D. viridis résiste aux attaques de l’oursin en évaluant

l’importance relative d’une défense chimique et mécanique.

Chapitre II

Temporal variation in community interfaces: kelp bed boundary dynamics adjacent to persistent urchin barrens

12

Résumé

Nous avons examiné, pendant deux ans, les facteurs qui affectent la stabilité

temporelle de la limite inférieure des ceintures de laminaires (Alaria esculenta) sur

cinq sites infralittoraux dans l’archipel de Mingan, dans le nord du golfe du

Saint‐Laurent. La position de la limite inférieure des ceintures variait beaucoup entre

les sites et dans le temps et était principalement contrôlée par l’oursin vert,

Strongylocentrotus droebachiensis, qui s’agrégeait en de denses (jusqu’à

500 individus m‐2) fronts de broutage en marge inférieure des ceintures. Ces fronts

avançaient sur les laminaires le plus rapidement pendant l’été (à des taux aussi élevés

que 2.5 m mois‐1), et il semblait y avoir une biomasse seuil d’oursins d’environ

5 kg m‐2 en‐dessous de laquelle les fronts ne pouvaient réduire substantiellement la

limite des ceintures. Les fronts étaient principalement composés de gros individus,

tandis que les plus petits oursins dominaient en zone dénudée, au‐dessous des

ceintures. À l’un des sites, nous avons noté de grands changements saisonniers dans

les densités globales d’oursins avec des augmentations et des baisses importantes

pendant l’été et l’hiver, respectivement. Une expérience comportant l’exclusion

d’oursins a indiqué que le recrutement algal en zone dénudée était de plus de cent

fois supérieur en absence qu’en présence d’oursins. La laminaire Agarum cribrosum

limitait grandement les mouvements de l’oursin et la stabilité la plus élevée d’une

ceinture sur l’un des sites semblait liée au remplacement graduel d’A. esculenta, dans

la partie inférieure de la ceinture, par une large bande d’A. cribrosum. Nous

proposons que des perturbations causées par des facteurs abiotiques (e.g., abrasion

par les glaces et mouvement de l’eau) déclenchent d’importants changements locaux

dans les densités d’oursins qui, à leur tour, déterminent principalement les limites de

distribution des ceintures de laminaires dans cette région de l’Atlantique où la zone

dénudée est un état communautaire dominant.

13

Abstract

We examined, over 2 years, factors affecting the temporal stability of the lower limit

of kelp beds (Alaria esculenta) at five subtidal sites in the Mingan Islands, northern

Gulf of St. Lawrence. The position of the lower limit of the beds varied markedly

among sites and over time and was largely controlled by the green sea urchin,

Strongylocentrotus droebachiensis, that formed dense (up to 500 individuals m‐2) feeding

fronts at the lower edge of the beds. These aggregations advanced over the kelp most

rapidly during the summer (at rates as high as 2.5 m month‐1), and there appeared to

be a threshold urchin biomass of ~5 kg m‐2 below which the fronts could not

substantially reduce the limit of the beds. The fronts consisted mainly of large

individuals, whereas smaller urchins predominated in the barrens zone below the

kelp beds. At one site, we recorded large seasonal shifts in overall urchin densities,

with large increases and decreases during the summer and winter, respectively.

An urchin exclusion experiment indicated that algal recruitment in the barrens was

two orders of magnitude greater in the absence than in the presence of urchins.

The kelp Agarum cribrosum greatly restricted urchin movements, and the greater

temporal stability of the kelp bed at one site appeared related to the gradual

replacement of A. esculenta in the lower kelp bed by a large stand of A. cribrosum.

We propose that perturbations by abiotic factors (e.g., ice scouring and water motion)

trigger important but localized changes in urchin densities which, in turn, largely

determine the limits of kelp bed distribution in this region of the Atlantic where

urchin barrens are a persistent community state.

14

Introduction

Kelp beds are highly diverse shallow coastal communities found in temperate and

polar zones worldwide. They are organized around the physical structure and

primary productivity provided by large brown algae of the order Laminariales

(Tegner & Dayton, 2000). In addition to being among the most productive systems on

earth (Mann, 1973), kelp beds provide habitat to many species for recruitment,

feeding, reproduction, and refuge from predators and from strong water motion

(e.g., Bernstein et al., 1981; Himmelman et al., 1983; Himmelman & Lavergne, 1985;

Duggins et al., 1990; Vadas & Elner, 1992; Vasquez, 1993; Balch & Scheibling, 2000;

Gagnon et al., 2003a). Although the light environment and substratum type are

primary factors determining where kelp beds develop, many other factors, such as

sedimentation, nutrients, salinity, water motion, and ice scour can modify local

patterns of occurrence (see review in Dayton, 1985b; but also Hooper, 1981;

Cowen et al., 1982; Ebeling et al., 1985; Keats et al., 1985; Seymour et al., 1989;

Sivertsen, 1997; Kawamata, 2001). The most important factor in many regions is

grazing by sea urchins (see review in Lawrence, 1975; but also Paine & Vadas, 1969;

Pearse & Hines, 1979; Hagen, 1983; Dean et al., 1984; Himmelman, 1984; Dayton,

1985a; Witman, 1987; Leinaas & Christie, 1996; Sivertsen, 1997;

Vasquez & Buschmann, 1997). Although urchins and kelp can co‐occur at small

spatial scales (i.e., within the foraging distance of urchins), high densities of urchins

usually completely denude areas of fleshy macrophytes, including kelp. Such areas,

referred to as urchin barrens (Lawrence, 1975), have both low diversity (but see

Bégin, 2002) and low productivity, although the abundance and diversity of coralline

algae can be quite high (Steneck, 1983; 1986; Chapman & Johnson, 1990;

Scheibling & Hatcher, 2001). Urchin barrens can be both extensive and persistent

(Breen & Mann, 1976b; Chapman, 1981; Hagen, 1983; Himmelman et al., 1983; Miller,

1985; Scheibling, 1986; Leinaas & Christie, 1996; Sivertsen, 1997), and natural and

experimental removals of urchins demonstrate that kelp can rapidly colonize barrens

when urchins are absent (e.g., Pearse & Hines, 1979; Himmelman et al., 1983;

Scheibling, 1986; Witman, 1987; Keats et al., 1990; Scheibling et al., 1999).

15

In the northwestern North Atlantic, the green sea urchin, Strongylocentrotus

droebachiensis, severely limits the distribution of kelp in rocky subtidal environments

and often forms dense feeding aggregations at the lower edge of kelp beds. This

phenomenon is reported in the Gulf of Maine (Sebens, 1985; Witman, 1985; 1987),

Newfoundland (Himmelman, 1984; Keats et al., 1985; Keats et al., 1990), and the

St. Lawrence Estuary (Himmelman et al., 1983; Himmelman & Lavergne, 1985).

In Nova Scotia (the most extensively studied region), urchin fronts can destroy kelp

beds (primarily Laminaria longicruris) at rates as high as 4 m month‐1, and as long as

the urchins remain sufficiently abundant, fleshy macroalgae cannot recolonize

(Breen & Mann, 1976a; b; Chapman, 1981; Miller, 1985; Scheibling, 1986; Scheibling et

al., 1999). As outlined by Scheibling et al. (1986; 1999), mass mortality of

S. droebachiensis caused by a pathogenic amoeba (see review in Scheibling, 1988) is the

only known natural mechanism for the large‐scale shift in community state from

urchin barrens to kelp beds. Such mortalities have been reported a number of times

over the last decades and were always followed by rapid kelp colonization

(Miller & Colodey, 1983; Scheibling & Stephenson, 1984; Miller, 1985; Scheibling,

1986; Scheibling & Hennigar, 1997; Scheibling et al., 1999). Although lobsters, crabs,

sea stars, wolffish, sea gulls and eider ducks are all predators of S. droebachiensis in the

northwestern North Atlantic (Himmelman & Steele, 1971), they have never been

shown to regulate urchin populations (see review in Scheibling, 1996).

Two such community states are found on shallow bedrock platforms in

wave‐exposed areas in the Mingan Islands, northern Gulf of St. Lawrence: (1) kelp

beds dominated by Alaria esculenta (but also including Laminaria spp., Saccorhiza

dermatodea, and Agarum cribrosum) in shallow water, and (2) urchin barrens

dominated by S. droebachiensis. Within the kelp bed, urchin densities are low, and

algal and invertebrate biomass is high (Bégin, 2002). In the barrens, urchins often

attain densities of 300 m‐2 (Himmelman & Nédélec, 1990; Gagnon et al., 2003b), and

the main algae are crustose coralline red algae (Lithothamnion sp. and Clathromorphum

sp.). The only common fleshy macroalgae in the barrens are the brown algae Agarum

cribrosum and Desmarestia viridis and the red alga Ptilota serrata. In most locations

16

A. cribrosum is found near the lower limit of the kelp zone where it usually forms

small patches (1‐10 m2), but it occasionally covers larger areas (>500 m2). Desmarestia

viridis, a filamentous annual, also forms patches in urchin barrens, but the patches are

typically small (1‐5 m2) and are only present from winter to early fall (Gagnon et al.,

2003b). Both A. cribrosum and D. viridis are thought to be chemically defended from

urchin grazing (Vadas, 1977; Larson et al., 1980; Himmelman & Nédélec, 1990;

Lobban & Harrison, 1994) although more recent work suggests that tactile contact

induced by water movement is more important, particularly for the finely‐branched

D. viridis (Gagnon et al., 2003b). Ptilota serrata often grows under patches of

A. cribrosum but can also form extensive turfs at greater depths.

Mass mortalities of urchins, as occur in Nova Scotia, has never been reported

in the northern Gulf. Thus, the urchin barrens appear to be a persistent community

state. The transition between kelp beds and barrens is distinct as the kelp generally

abruptly disappear where fronts of urchins begin (Fig. 2.1). The factors controlling the

stability of this conspicuous interface are poorly understood. Qualitative observations

over 20 years (Himmelman and coworkers) suggest that its position (i.e., its depth

and distance from shore) varies over time from site to site (with wave exposure and

slope). Although the kelp‐barrens interface represents an equilibrium in an important

herbivore‐plant interaction, there are no long‐term observations of how it fluctuates

over time, or an understanding of how such changes are affected by other

environmental factors (e.g., ice scour, dislodgment by waves and currents,

competition). A necessary first step is a detailed examination of the spatial and

temporal dynamics of this interface relative to the demographic changes in both the

kelp bed and the urchin population.

Here, we examine factors affecting the temporal stability of the lower limit of

kelp beds in the Mingan Islands. Specifically, we (1) record temporal variations in the

position of the lower limit of kelp beds, (2) measure temporal changes in the structure

(density and size‐distribution) of kelp beds and urchin fronts in the transition zone,

17

Fig. 2.1. A grazing front of the green sea urchin, Strongylocentrotus droebachiensis (2 to 3 layers thick), at a depth of 3‐4 m at the lower edge of a subtidal kelp (Alaria esculenta) bed in the Mingan Islands in June 2000. The urchins are actively consuming kelp blades and leaving only the bases of stipes in their path.

18

(3) examine the relationship between urchin biomass in grazing fronts and the rate

of change in the position of the lower limit of kelp beds, (4) determine the impact

of urchins on the recruitment of kelp in the barrens zone, and (5) assess the

effectiveness of A. cribrosum stands in limiting upward movements of urchins into the

shallower kelp zone dominated by A. esculenta.

Materials and methods

Study sites

Our study was conducted during the summer and fall of 2000, 2001, and 2002 at

five subtidal sites on four of the Mingan Islands, northern Gulf of St. Lawrence

(Fig. 2.2). All sites were gently‐sloping bedrock platforms and had kelp beds in

shallow water (to a maximal depth of 6 m) and urchin barrens in deeper water. At the

beginning of the study, the kelp beds at the sites were composed primarily of

Alaria esculenta and the transition from kelp beds to barrens was abrupt. The sites

varied markedly in exposure to waves and tidal currents due to differences in

orientation and topography. After initial qualitative surveys, two sites, Île à Firmin

East and Île du Havre, were selected for intensive study due to striking differences in

algal density within the kelp beds (kelp were more dense at the former site).

The Île à Firmin East site was exposed to strong wave action and tidal currents,

whereas the Île du Havre site had less exposure to waves and had moderate tidal

currents. The three other sites, Île Niapiskau, Île à Firmin West, and Petite île au

Marteau, were studied less intensively and served to provide information on the

generality of the temporal patterns observed at the two intensively studied sites.

We did not make measurements during other seasons due to the remote location of

the field sites, unpredictable weather conditions, and ice cover in winter.

Temporal variability in A. esculenta beds and urchin populations

We evaluated community shifts by monitoring changes over 2 years in (1) the

19

Fig. 2.2. Location of study sites at Île Niapiskau, Île à Firmin (West and East), Île du Havre, and Petite île au Marteau, in the Mingan Islands, northern Gulf of St. Lawrence, eastern Canada.

20

absolute position of the lower limit of the kelp bed, and (2) the density and

size‐distribution of kelp and urchins in the kelp‐barrens transition zone. For each

of the two intensively studied sites, we delineated 12 permanent 16‐m‐long transects

that ran perpendicular to the kelp‐barrens transition and parallel to one another.

The transects were spaced at 3‐m intervals and thus covered 33 m of shoreline.

The ends of each transect were tied to two 33‐m‐long benchmark lines (one located in

the kelp bed and the other in the barrens) running parallel to the shoreline and

anchored by bolts (3 on each line) permanently set into holes drilled into the

substratum. The transect lines were removed between sampling periods.

The absolute position of the kelp‐barrens interface was determined by

measuring the distance (to the nearest 1 cm) from the deep end of each transect up to

the interface. To measure the abundance and size of urchins and kelp, we placed

quadrats along each transect line in both seaward and landward directions from the

kelp‐barrens interface. Quadrats were positioned by placing the edge of a quadrat

(25 x 25 cm) at distances of 0.0 (i.e., at the interface), 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0 and

5.0 m in each direction (this arrangement placed the two quadrats at 0.0 m

immediately adjacent to one another). In each quadrat we counted the number of

urchins in three size classes (<2 cm, 2‐4 cm, and >4 cm in test diameter). We also

categorized each quadrat according to the density of kelp (0‐10, 10‐30, 30‐50,

and >50 individuals per quadrat) and their length (<50 cm, 50‐100 cm, and >100 cm).

Sampling was done six times over a 2‐year period: June, August, and October 2000,

June and August 2001, and June 2002. We could not, however, take all measurements

each time. Bad weather precluded the measurement of the density and

size‐distribution of kelp and urchins in October 2000, and time considerations limited

the quadrat samplings in 2001 and 2002 to 5 distances (0.0, 0.5, 1.0, 3.0, and 5.0 m) in

each direction from the kelp‐barrens interface.

At the other three sites, we simply measured the absolute position of the

kelp‐barrens interface along 20 m of shoreline. At each site, a transect line was placed

along the lower limit of the kelp bed (two positions were permanently marked

21

by bolts set into the substratum), and on each observation date (August 2000,

June and August 2001, and June 2002), we recorded the perpendicular distance to the

lower limit of the kelp bed at 5‐m intervals along the transect line. The line was

removed between sampling dates.

Should grazing by urchins cause the reduction of kelp beds, the rate of kelp

bed regression should be positively correlated with urchin biomass in the grazing

fronts. To test this hypothesis, we examined the data on the position of the lower limit

of kelp beds and urchin densities acquired from the 2‐year monitoring at Île à Firmin

East and Île du Havre sites. For each site, we first determined the monthly rate

of change (advance or regression) in the mean position of the lower limit of the kelp

bed (the average over the twelve transect lines) by comparing their position at the

beginning and at the end of each of four time periods (between June and August 2000

and 2001, and between August and June 2000‐2001 and 2001‐2002). For each period,

we then calculated the urchin biomass within the first meter of the edge of the

barrens (i.e., the front). We used the average of the quadrats placed at 0.0, 0.5, and

1.0 m from the edge and converted densities to biomass using a regression

(Wet weight (g) = 0.000908 x Diameter (mm)2.804; r2 = 0.99) obtained from a sample of

144 urchins (ranging in test diameter from 0.8 to 6.1 cm) collected at the Île du Havre

site. Test diameters used for each size class (<2 cm, 2‐4 cm, and >4 cm) were 1, 3, and

5 cm, respectively, and densities were the average of the urchins sampled at the

beginning and end of each period. We then examined the relationship between

urchin biomass in the grazing fronts and monthly rate of changes in the position

of the lower limit of kelp beds at each site.

Effect of urchins on algal recruitment in the barrens

To quantify the impact of urchins on algal recruitment in the barrens,

we experimentally manipulated urchins using exclosures. The experiment consisted

of three treatments: 1. Grazed (no manipulation) to quantify algal recruitment under

natural urchin densities; 2. Ungrazed (exclosure and urchins manually removed)

to quantify algal recruitment in the absence of urchins; and 3. Fence control

22

(exclosure, but urchins not manually removed) to test for effects of the exclosure

beyond those associated with the presence of grazers.

For the Ungrazed and Fence control treatments, we constructed exclosures that

consisted of a 0.75 x 0.75‐m fence of 60‐cm‐wide strips of hardware cloth (1‐cm mesh

size) supported by a PVC frame. The fences were anchored to the substratum with

nylon ropes connecting the top of each corner to eye‐bolts set into the bedrock.

Urchins were able to climb the fences but with visits every 2‐3 d to hand remove

them, we were able to limit greatly their penetration inside the fences. Fence controls

were identical except that urchins were never removed after starting the experiment.

We selected 12 experimental areas (each marked with a bolt in the center) in

June 2000, spaced at 3.8‐m intervals 3 to 4 m below the lower limit of the kelp bed at

Île du Havre (near the transects at this location). The areas were grouped into four

blocks of three adjacent areas that were then randomly assigned the three treatments.

The experiment began on 3 July 2000, when we removed the urchins from all the

areas (releasing them at least 3 m from the areas) and then wire‐brushed the rock

surface to remove any fleshy macroalgae present (e.g., rare recruits of Laminaria sp.

and A. cribrosum), thereby increasing the homogeneity of the initial conditions of the

experimental units. The experiment ended 44 d later at which time we quantified the

density of kelp recruits in each area using two quadrats (25 x 25 cm) placed on the

north and south sides of the center bolt in each area. We averaged the data from the

two quadrats in each area before statistically comparing densities of kelp recruits

among treatments.

Limitation of upward movements of urchins by A. cribrosum

During the first year of the study, changes were seen in the kelp‐barrens transition

that were related to an observed increase in the abundance of Agarum cribrosum at the

Île du Havre site. To assess the effectiveness of mature stands of A. cribrosum in

limiting the upward movement of urchin fronts into the kelp (Alaria esculenta) zone,

we conducted an experiment to compare the rate of penetration of urchins in the

absence and presence of adult A. cribrosum. The two treatments consisted of

23

A. cribrosum removal (stipes cut off at the holdfasts) and a control (no manipulation).

Experimental units consisted of eight 4‐m2 areas spaced at 3.8‐m intervals along the

lower edge of a large (over 500 m2) mature stand of A. cribrosum located near the

transect lines at the Île du Havre site. Each area, 1‐m wide and extending 4 m up into

the A. cribrosum zone, was marked with bolts at the lower and upper edge. The areas

were grouped into four blocks of two adjacent areas to which the two treatments

were randomly assigned. The experiment began on 27 June 2001 with the removal of

all adult and juvenile A. cribrosum for the removal treatment and the removal of all

urchins (released >25 m away) from all experimental units. The movement of urchins

into the experimental units was then recorded on 6 dates over the next 20 d (June 28,

29 and 30 and July 4, 10 and 17). On each date we counted the number of urchins in

three size classes (<2 cm, 2‐4 cm, and >4 cm) in quadrats (40 x 40 cm) placed at 0, 1, 2,

and 3 m from the edge of the original A. cribrosum zone in each area. To provide an

indication of the overall urchin densities, we then averaged the densities of urchins

from the 1‐, 2‐ and 3‐m distances in each area on each date before comparing

treatments.

Statistical analysis

Temporal changes in the density of sea urchins in the kelp‐barrens transition zones

of Île à Firmin East and Île du Havre were analyzed using a repeated measures

ANOVA (Hand & Taylor, 1987; Crowder & Hand, 1990) with the factors, Time

(June 2000, July 2000, June 2001, July 2001, and June 2002) and Site (Île à Firmin East

and Île du Havre) with one nested element (transect replicate within island). Since

variances did not vary over time within the islands, and the correlation in the data

over time was similar between the islands, we used a pooled covariance structure in

the ANOVA (Proc Mixed, Type = CS, SAS Institute Inc., 1999). We applied the

analysis to the raw data as the data for each time were normally distributed.

We examined the relationship between urchin biomass in grazing fronts and

rate of change in the position of the lower limit of kelp beds at Île à Firmin East and

Île du Havre using regression analysis (Draper & Smith, 1981; Proc Reg, SAS Institute

24

Inc., 1999). We applied the analysis to the raw data as the data for each site were

homoscedastic and normally distributed. We evaluated the effect of urchins on algal

recruitment in the barrens using a two‐way ANOVA (Zar, 1999) with the factors,

Block (four) and Treatment (Grazed, Ungrazed, Exclosure control) applied to the raw

data as they were homoscedastic and normally distributed. To test whether

Agarum cribrosum impeded the upward movement of urchin fronts, we carried out a

repeated measures ANOVA with the factors, Block (four), Treatment (with/without

canopy), and Time (after 1, 2, 3, 7, 13, and 20 d) with one nested element (treatment

within block). Since we detected fluctuations in the variances over time, we modeled

the covariance structure with a heterogeneous compound symmetry structure

(Proc Mixed, Type = CSH, SAS Institute Inc., 1999). We analyzed the raw data as the

data for each time were normally distributed.

In the above analyses, normality was verified using Shapiro‐Wilk’s statistic

(SAS Institute Inc., 1999) and homoscedasticity by examining the graphical

distribution of the residuals and by applying Levene test (Snedecor & Cochran, 1989).

To detect differences among levels within a factor, we used least‐square means

multiple comparisons tests (LS means, SAS Institute Inc., 1999). A significance level

of 0.05 was used for all statistical tests.

Results

Temporal variability in the lower limit of kelp beds

During the 2‐year study, the lower limit of kelp beds showed greater temporal

variability at Île à Firmin East than at Île du Havre (Fig. 2.3). At Île à Firmin East, the

kelp bed retreated steadily during the first year and by August 2001 had been pushed

back by 11.9 ± 5.0 (SD) m (equivalent to a loss of 0.9 m month‐1). Nearly 40 % of this

loss occurred between June and August 2001 (2.5 m month‐1). There was a strong

recruitment event during the winter of 2001‐2002, and the kelp bed reinvaded

a distance of 13.7 ± 4.8 m, exceeding by 1.8 ± 2.0 m the initial lower limit. In contrast,

25

0

10

20

30

40Jun 2000 Aug 2000

Aug 2001 Jun 2002

Île à Firmin EastOct 2000

0

10

20

30

40

Île du HavreJun 2000 Aug 2000 Oct 2000

0 6 12 18 24 30

Aug 2001

0 6 12 18 24 30

Jun 2002

Alongshore distance (m)

0

10

20

30

40Jun 2001

0 6 12 18 24 300

10

20

30

40Jun 2001

Barrens

Kelp bed

(8.1±1.9) (11.6±1.9) (13.7±1.6)

(15.6±4.6) (20.5±4.2) (6.7±1.7)

(8.7±1.5) (9.4±1.9)

(9.6±1.9) (9.1±2.2)

(12.1±1.2)

(8.2±2.7)

0

0

0

0

Fig. 2.3. Changes in the position of the kelp‐barrens interface at Île à Firmin East and Île du Havre between June 2000 and June 2002. The depth across the grid (from 40 to 0 m along the y‐axis) was from 3 to 5 m at Île à Firmin East, and from 4 to 6 m at Île du Havre. Values in parenthesis represent the mean distance (± SD) of the kelp interface relative to the benchmark transect in the urchin barrens (0 m). Horizontal dashed lines indicate the mean distance of the kelp bed on the first sampling date (June 2000).

26

the kelp‐barrens interface at Île du Havre was relatively stable over the 2‐year period:

the greatest regression was just 3.5 ± 1.3 m, with a maximal rate of loss

(1.3 m month‐1) between August and October 2000. Some recolonization occurred

during the first winter so that the lower limit in June 2001 was only 0.9 ± 1.7 m above

the lower limit of the bed in the previous June. The bed extended deeper during the

second year, and, as for the Île à Firmin East site, slightly exceeded the initial lower

limit by the end of the study. Thus, although monthly changes in the position of the

lower limit of the kelp bed were much greater at Île à Firmin East than at Île du

Havre, the net displacement of the lower limit of the kelp bed over 2 years was

similar at the two sites.

Measurements of the lower limit of the kelp beds at the three less‐intensely

studied sites also showed a high degree of variability among sites and over time. At

two sites, Île Niapiskau and Petite île au Marteau, there was a substantial regression

of the kelp bed (over a distance of 3.8 ± 1.1 m and 5.9 ± 1.4 m, respectively) during the

first 12 months, as observed at Île à Firmin East. In the second year, there was little

change at Petite île au Marteau as the lower limit of the bed in June 2002 was only

1.0 ± 1.3 m below the lower limit in August 2001. However, as seen at Île à Firmin

East, substantial recolonization appeared to occur at Île Niapiskau during the winter

of 2001‐2002. We were unable to locate reference points at this site in June 2002 due to

the kelp cover, but noted that the lower limit of the kelp bed was well below that of

the previous summer. In contrast to all other sites, the third site, Île à Firmin West,

showed a major extension of the kelp bed during the first winter (7.1 ± 1.2 m) and

then retreated during the summer season (4.4 ± 1.9 m). We were also unable to find

the reference points at this site in 2002, but again observed another major extension of

the kelp bed during the second winter. There appeared to have been a major

disturbance over the winter 2001‐2002 as cobbles and boulders covered with small

kelp were scattered over the bottom. Also, there were denuded areas at shallower

depths in the kelp bed that were likely due to ice scour. None of these sites showed

the stability in the kelp‐barrens interface seen at the Île du Havre site.

27

Structure of kelp beds

The 2‐year monitoring of the density and size of kelp in the transition zone indicated

that kelp in the lower 5 m of the bed were always more abundant (Table 2.1) and

larger (Table 2.2) at Île à Firmin East than at Île du Havre. At Île à Firmin East, there

were >50 kelp plants in almost half of the quadrats at the beginning of the study.

However, densities decreased sharply during the summers of 2000 and 2001 as the

kelp bed retreated to a shallower depth (Fig. 2.3). The major increase in June 2002

relative to August 2001 was due to the newly recruited kelp that caused a major

extension of the kelp bed (Fig. 2.3). The occurrence of >100‐cm plants was relatively

high at all times (13‐32 %). Alaria esculenta always dominated, but Laminaria digitata

and Saccorhiza dermatodea were also common and patches of the green alga

Spongomorpha arcta, ulvoid algae, and the brown alga Scytosiphon lomentaria were

present on surfaces between kelp holdfasts, especially in June of each year.

In contrast, at Île du Havre, there were <10 kelp plants in most quadrats at the

beginning of the study. Densities decreased further over time and were extremely

low by June 2002. This decrease in kelp density was accompanied by a general

decrease in kelp size. For example, the occurrence of plants >50 cm in length was

initially only 29 % but decreased to 5 % by the end of the study. The decline in both

kelp density and length was due to a shift from A. esculenta to Agarum cribrosum in the

lower kelp bed (in mature stands, A. cribrosum rarely exceeds 1 m in length, and

individuals are more dispersed than in a bed of A. esculenta; P. Gagnon, personal

observation). By June 2002, the lowest portion of the bed was a mature stand

of A. cribrosum that extended up to 8 m from the lower limit of the A. esculenta zone.

Above the A. cribrosum stand, a luxuriant kelp bed of large A. esculenta still flourished.

There were also a few large Laminaria longicruris (often >2 m in length), but L. digitata

and S. dermatodea were virtually absent, and S. arcta, ulvoid algae, and S. lomentaria

were scarce. We never observed any major recruitment events at this site.

28

Table 2.1. Proportion of quadrats (out of 48) containing different density classes of kelp (0‐10, 10‐30, 30‐50, and >50 individuals per 25 x 25‐cm quadrat) on different dates from June 2000 to June 2002 for the lower 5 m of the kelp bed zone (quadrats placed at 0.0, 1.0, 3.0 and 5.0 m) at Île à Firmin East and Île du Havre.

Île à Firmin East Île du Havre

0‐10 10‐30 30‐50 >50 0‐10 10‐30 30‐50 >50

Jun 2000 27 12 15 46 61 35 0 4

Aug 2000 54 27 8 11 90 10 0 0

Jun 2001 50 17 21 12 92 4 2 2

Aug 2001 94 6 0 0 98 2 0 0

Jun 2002 21 19 27 33 100 0 0 0

29

Table 2.2. Proportion of quadrats (out of 48 25 x 25‐cm quadrats) containing different size classes of kelp (<50, 50‐100, and >100 cm in length) on different dates from June 2000 to June 2002 for the lower 5 m of the kelp bed zone (quadrats placed at 0.0, 1.0, 3.0 and 5.0 m) at Île à Firmin East and Île du Havre.


<50 cm 50‐100 cm >100 cm <50 cm 50‐100 cm >100 cm

Jun 2000 73 13 14 71 21 8

Aug 2000 73 12 15 82 10 8

Jun 2001 74 4 22 97 1 2

Aug 2001 58 10 32 91 5 4

Jun 2002 51 36 13 95 5 0

30

Structure of urchin populations

The 2‐year monitoring of the density and size‐structure of urchin populations in the

transition zone showed substantial changes at each site as well as major differences

between sites. There was an overall decline in urchin abundance by approximately

50 % during the course of the study at both sites (Fig. 2.4). At Île à Firmin East, this

trend was seasonally reversed with dramatic increases during the summer. In

contrast, no seasonal variation was seen at Île du Havre, and a massive decline in

densities (47 %) between August 2000 and June 2001 accounted for most of the

change during the study.

The sampling at different distances from the kelp‐barrens interface (into the

barrens and into the kelp bed) showed that urchin densities consistently peaked in

the first meter below the edge of the bed at both sites (Fig. 2.5). There was almost

always a distinct front of urchins, often 2‐3 layers thick. Urchin abundance fell off

much more rapidly within the kelp bed than in the urchin barrens although at times,

there were substantial densities (>100 m‐2) within the first 2 m, and occasionally even

far into the kelp bed (e.g., Île du Havre in June 2001). The bell‐shaped distribution

around the kelp‐barrens interface flattened with time, particularly at Île à Firmin East,

associated with the overall decrease in urchin numbers (see above).

The analysis of the size‐structure of urchins in the transition zone showed that

mean size generally increased over time at both sites, but urchins were always larger

at Île à Firmin East (Fig. 2.6). In the kelp bed above the interface, urchins were

generally larger with the largest size class often comprising more than 50 % of the

sample and the smallest size class usually accounting for <25 %. Most urchins in the

front and in the kelp bed were medium (2‐4 cm) to large (>4 cm) in size (Fig. 2.6)

whereas small urchins (<2 cm) were far more common in the urchin barrens,

particularly at Île du Havre. At Île à Firmin East, the proportion of small urchins in

the overall area decreased during the first year (to nearly nil in the summer of 2001)

and then increased by June 2002. Urchin biomass in the grazing front (in the first

31

0

50

100

150

200

250 Île à Firmin East

Île du Havre

Jun 00 Aug 00 Jun 01 Aug 01 Jun 02

abaa

bc

cdde

dede e

f

Fig. 2.4. Changes in the mean density (+ SE) of the green sea urchin, Strongylocentrotus droebachiensis, within 5 m of the lower limit of kelp beds at Île à Firmin East and Île du Havre between June 2000 and June 2002. Values not sharing the same letter are different (LS means, P < 0.05).

32

0100200300400500600


Distance to the kelp bed edge (m)

0100200300400500600

0100200300400500600

0100200300400500600

0100200300400500600

Jun 2000

Aug 2000

Jun 2001

Aug 2001

Jun 2002

Jun 2000

Aug 2000

Jun 2001

Aug 2001

Jun 2002

Kelp bed Barrens Kelp bed Barrens

Edge Edge

2 222 4 66 644 6 4

Fig. 2.5. Changes in the densities of the green sea urchin, Strongylocentrotus droebachiensis, within 5 m of the lower limit of kelp beds at Île à Firmin East and Île du Havre between June 2000 and June 2002. Data are means (+ SE) for quadrats (n = 12) placed at different distances inside and outside of the kelp‐barrens interface (0.0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0 and 5.0 m in June and August 2000, and at 0.0, 0.5, 1.0, 3.0, and 5.0 m in June and August 2001, and June 2002).

33

Medium

Large

Jun 2000

Aug 2000

Île à Firmin East

Jun 2001

Aug 2001

Jun 2002

Île du Havre

Distance to the kelp bed edge (m)

Small

Kelp bed Barrens0-1 3 535

Jun 2000

Aug 2000

Jun 2001

Aug 2001

Jun 2002

1-0

5.5 2.0 2.5

7.7 3.6 2.7

7.6 2.7 5.1

10.2 5.5 6.3

3.2 3.3 2.4

5.6 2.8 1.7

3.7 3.6 1.4

4.5 2.3 2.6

4.6 3.1 2.2

4.8 2.9 2.5

0.6 0.5 3.6

1.1 3.5 5.9

0.7 0.1 2.7

0.7 3.2 5.2

0.2 0.4 0.5

1.8 1.0 2.6

2.5 3.4 2.8

6.2 3.5 1.8

3.2 2.4 2.3

0.8 3.3 1.2

Edge

Fig. 2.6. Changes in the proportions of small (<2 cm test diameter), medium‐sized (2‐4 cm) and large (>4 cm) urchins, Strongylocentrotus droebachiensis, within 5 m of the lower limit of kelp beds at Île à Firmin East and Île du Havre between June 2000 and June 2002. Data are means for quadrats (n = 12) placed at 0.0, 0.5, and 1.0 m (averaged and designated as 0‐1 m interval), and at 3.0 and 5.0 m inside and outside of the kelp‐barrens interface. Values above the circle plots represent urchin biomass (kg m‐2) estimated from a regression equation relating fresh weight to test diameter (n = 144) collected at Île du Havre.

34

meter below the kelp edge) was initially similar (∼5.5 kg m‐2) at the two sites (Fig. 2.6).

Although the biomass increased almost twofold over the following 14 months at

Île à Firmin East (to August 2001), it always remained less than the initial value

at Île du Havre. Thus, during most of the study, the grazing potential was higher at

Île à Firmin East than at Île du Havre.

Our data did not support the hypothesis that the rate of kelp bed regression is

positively correlated with urchin biomass in the grazing fronts. We found no

significant relationship between the biomass of urchins in the grazing fronts and the

rate of change in the position of the lower limit of kelp beds at the two intensively

studied sites, either when averaged for the entire site (Fig. 2.7) or when examined

within a site using transects as replicates (data not shown). Nevertheless, we noted a

tendency towards increased kelp regression with increased urchin biomass at

Île à Firmin East (r2 = 0.33, P = 0.42). There appeared to be a threshold biomass of

urchins of ~5 kg m‐2 (estimated from a regression equation of the pooled data from

both sites; i.e, when kelp edge displacement (m) = 0 = 1.79 ‐ 0.37 x Urchin biomass

(kg m‐2), r2 = 0.29, P = 0.17) above which the beds receded. As urchin biomass was

usually just below 5 kg m‐2 (June 2000 being the only exception) at Île du Havre, this

might explain why changes in the position of the kelp bed interface varied little

during the study (but see below). Overall, these results indicate that the variability in

the position of the kelp‐barrens interface is a more complex function than simply

being related to the urchin biomass in the grazing front (e.g., related to kelp density,

wave exposure, tendency of urchins to feed).

Effect of urchins on algal recruitment in the barrens

The 44‐d exclusion of urchins from experimental areas in the urchin barrens

demonstrated that urchins strongly limited kelp recruitment as recruitment was

>100 times greater in the Ungrazed treatment (232.0 ± 66.5 [SE] recruits m‐2) than in

the Grazed treatment (2.0 ± 2.0; LS means, P = 0.0053). Early colonists, such as

diatoms, ulvoid algae, and Spongomorpha arcta, also flourished in the absence of

35

-3

-2

-1

0

1

2

0 2 4 6 8

Urchin biomass (kg m-2)

10

Fig. 2.7. Relationship between the rate of change in the position of the lower limit of kelp beds and urchin biomass in the grazing fronts at Île à Firmin East (filled circles) and Île du Havre (open circles) between June 2000 and June 2002. Positive rates indicate kelp bed expansion and negative rates kelp bed regression.

36

urchins. The similarity between the Fence control (4.0 ± 2.3) and Grazed treatments

(LS means, P = 0.97) demonstrated that the manipulation did not affect recruitment

beyond the effect of grazing by urchins.

Limitation of upward movements of urchins by A. cribrosum

The experimental removal of mature Agarum cribrosum showed that dense stands of

kelp can greatly reduce shoreward movement of urchins. In the areas where

A. cribrosum was not removed, the density of urchins in the lower A. cribrosum zone

increased gradually over the first 3 d of the experiment but then remained low for the

duration of the experiment (<34 m‐2, Fig. 2.8). In contrast, in the areas where the

A. cribrosum canopy was removed, the density of urchins increased (particularly

between 3 and 7 d), and after 20 d was more than four times greater than the first day.

The final urchin density (after 20 d) was three times greater in the removal than in the

non‐removal areas. During the experiment, we observed dense urchin aggregations at

the outer edge of the intact A. cribrosum areas (up to 356 urchins m‐2, SE = 177.1)

during periods of calm wave conditions. Conditions were especially calm during the

last 3 d of the study, and we observed groups of urchins at the edge of the bed

anchoring the blades of A. cribrosum to the bottom. These aggregations likely

accounted for the large increase in the density of urchins in the intact A. cribrosum

areas as recorded on Day 20.

Discussion

Although large seaweeds generally dominate shallow rocky subtidal communities on

northwestern North Atlantic coasts (Chapman & Johnson, 1990) grazing by the green

sea urchin, Strongylocentrotus droebachiensis, nevertheless severely limits the

distribution of kelp (Himmelman & Steele, 1971; Breen & Mann, 1976a; b;

Chapman, 1981; Himmelman et al., 1983; Himmelman, 1984; Miller, 1985; Scheibling,

1986; Witman, 1987; Keats et al., 1990; Scheibling et al., 1999). Our urchin exclusion

experiments showing that urchins dramatically limited the recruitment of both kelp

37

a

a

a

bbbc

ccdcd

d

de

e

0

20

40

60

80

100

120

140

Intact A. cribrosum

A. cribrosum removed

Time (d)

1 2 3 7 13 20

Fig. 2.8. Changes, over a 20‐d period, in the mean density (+ SE) of the green sea urchin, Strongylocentrotus droebachiensis, in 1 x 4‐m areas along the lower edge of a mature stand of the kelp Agarum cribrosum where the kelp was removed and in 1 x 4‐m areas where A. cribrosum was left intact. Values not sharing the same letter are different (LS means, P < 0.05).

38

and ephemeral algae in the barrens is consistent with this body of knowledge. Thus,

as seen elsewhere, urchin grazing is a primary factor controlling algal distribution in

the shallow rocky subtidal zone of the Mingan Islands.

Most previous studies have focused on the negative impact of the green sea

urchin on primary productivity, but these one‐time observations have only provided

limited information about interactions controlling changes in kelp bed boundaries

(but see Scheibling et al., 1999). Here we have shown that the kelp‐barrens interface is

dynamic and can shift over weeks to months. Moreover, we have demonstrated that

these changes in the lower limit of kelp beds are site‐specific. For example, between

August 2000 and June 2001, three sites experienced a regression in the kelp‐barrens

interface of 4 to 7 m, one site experienced no change, and one site showed an advance

of 7 m. This variation was not clinal as the two most extreme sites (Île à Firmin East

and West) were the closest together. Substantial temporal variation was also seen,

especially at Île à Firmin East where there was a continuous regression of the

kelp bed from 2000 to 2001, followed by massive kelp recolonization in 2002.

We had hypothesized that changes in the interface would be related to changes

in the urchin population. Whereas the massive recruitment at Île à Firmin East

occurred when urchin densities were at their lowest (August 2001 to June 2002), the

movement of the interface did not track variations in urchin densities. For example,

in spite of a 44 % decrease in urchin density during the winter of 2000‐2001, the kelp

bed continued to regress. However, rates of regression were always highest during

the summer (up to 2.5 m month‐1), when urchin numbers were increasing. Because

urchins feeding rates typically peak in summer (Vadas, 1977; Larson et al., 1980;

Himmelman, 1984; Himmelman & Nédélec, 1990; Meidel & Scheibling, 1998), their

increased voraciousness likely enhanced the effect of their increased density.

The rates of kelp bed regression that we observed were similar to those seen along the

coast of Nova Scotia (up to 4.0 m month‐1, Breen & Mann, 1976b; Mann, 1977;

Scheibling et al., 1999). Overall, our data suggest that the effect of urchins was

non‐linear at Île à Firmin East and that kelp could only colonize when the urchin

39

density was below ~100 urchins m‐2. Possibly, the low number of urchins reduced

their ability to weigh down and graze on the kelp. Biomass offers another measure of

grazing pressure, and Breen and Mann (1976a; b; and later supported by Scheibling

et al. 1999) postulated that a threshold urchin biomass of ~2 kg m‐2 causes kelp

regression in Nova Scotia, a value less than half the threshold that we estimated for

the Mingan Islands (~5 kg m‐2). The difference might be attributable to the higher kelp

density in the Mingan Islands, or to the morphology or chemistry of the dominant

kelps in the two systems, Laminaria longicruris in Nova Scotia and Alaria esculenta in

the Mingan Islands. Another possibility is that our sites were more exposed to wave

action or tidal currents than the sites in Nova Scotia, as increased wave action likely

decreases the rate at which urchins move and attack algae (Kawamata, 1998;

Chapter 5). The greater exposure of the site at Île à Firmin East to waves and currents,

compared to the site at Île du Havre, could have contributed to the increased

variability in the lower limit of the kelp bed at this site. If urchins at Île à Firmin East

experienced less predictable flow forces, they may have been able to move rapidly

upwards during periods of low wave action. Likewise, unexpected periods of high

wave action could have dislodged urchins and permitted a massive expansion of

the kelp bed. Finally, perhaps neither density nor biomass are individually the best

measure of grazing pressure, and the best may be some combination of the two.

The large decrease in urchin density at Île à Firmin East during the winter

of 2001‐2002 was accompanied by an increased proportion of small urchins. Because

of the much lower ingestion rate of small compared to large urchins (the rate for

<15‐mm urchins is 10‐fold less than for 50‐mm urchins, Larson et al., 1980;

Himmelman, 1984; Minor & Scheibling, 1997; Meidel & Scheibling, 1999;

Scheibling & Hatcher, 2001), the overall grazing pressure may have been too low to

prevent colonization by kelp. Several observations from Île du Havre support this

idea. First, urchin densities were always >100 m‐2 yet no massive colonization was

observed (nevertheless, the greatest extension of the kelp bed did occur when urchin

densities declined by 47 % in the winter of 2000‐2001). On the other hand, in spite of

the constant high urchin densities, there was never a substantial regression of the

40

kelp bed, even though kelp density was always lower and the plants smaller in the

lower kelp bed compared to Île à Firmin East, which should have facilitated the

advance of the front (Scheibling et al., 1999). Our data support the hypothesis of a

threshold urchin biomass as the grazing front at Île du Havre remained below

5 kg m‐2 throughout most of the study and was apparently insufficient to cause

a substantial regression of the interface there.

Another major difference between the two sites was the difference in kelps that

formed the beds. Alaria esculenta was the dominant kelp at Île à Firmin East, whereas

we observed a shift from A. esculenta to Agarum cribrosum at the lower edge of the bed

at Île du Havre during the study as reflected by the gradual decline in mean density

and height of the kelp. Agarum cribrosum is probably more resistant to grazing than

A. esculenta given the low preference of the urchin for this alga in both field and

laboratory feeding studies (Himmelman & Steele, 1971; Vadas, 1977; Larson et al.,

1980; Himmelman, 1984; Himmelman & Nédélec, 1990). This low preference may be

partially due to chemical defenses, such as phenolics (Vadas, 1977; Larson et al., 1980;

Himmelman & Nédélec, 1990). The tough stipe of A. cribrosum, which is fairly rigid

and holds the blade above the bottom, may also contribute to its resistance to grazing.

The greater resistance of A. cribrosum compared to A. esculenta is further shown by its

ability to form patches within the barrens (Vadas, 1968; 1977; Keats et al., 1982;

Himmelman, 1984). Moreover, as shown by our A. cribrosum removal trials, the

presence of A. cribrosum can greatly reduce the upward movement of urchins. Thus,

in certain situations, A. cribrosum appears to form a border at the lower edge of the

kelp zone that limits incursions by urchins and in turn protects the more vulnerable

A. esculenta growing at shallower depths from urchins. This effect on urchins may be

unique to A. cribrosum because of its distinct grazing resistant characteristics but

could also be a general characteristic of any alga capable of forming a dense stand.

The mechanism by which A. cribrosum reduces movements of urchins remains

to be determined. The alga might ”block” the movement of urchins, by forming

a barrier or repulsing the urchins, as seen for some other macroalgae

41

(Velimirov & Griffiths, 1979; Himmelman, 1984; Dayton, 1985a; Vasquez & McPeak,

1998; Konar, 2000; Gagnon et al., 2003b), or urchins might simply stop once they

encounter a potential food source. A slowing of the advance of urchins by

A. cribrosum was suggested by our observations at Île du Havre. Although the urchin

front was clearly defined in June 2000 (high numbers at the lower edge of the kelp

bed), the front became less pronounced over time as A. cribrosum gradually replaced

A. esculenta. Because most of the urchins aggregated at the lower edge of the

A. cribrosum, fewer urchins penetrated to the A. esculenta in shallower water. The

greater temporal variability in the position of the lower limit of the kelp bed at

Île à Firmin East, compared to Île du Havre, may thus have been related to the

absence of A. cribrosum at this site.

We observed substantial declines in urchin abundance over time at both Île du

Havre and Île à Firmin East, but events varied between sites. At Île du Havre,

densities declined abruptly between August 2000 and June 2001, coincident with an

increase in the average size, suggesting a demographic change (such as mortality or

emigration). In contrast, at Île à Firmin East, we observed a more gradual temporal

decline along with a seasonal pattern. The seasonal pattern was likely due to the

migration of urchins into deeper water during the winter, perhaps to avoid ice scour

(see below) or increased wave action as suggested for urchins populations in the

St. Lawrence Estuary (Himmelman & Lavergne, 1985), and their return the following

spring. The gradual decline in density by >50 % between June 2000 and June 2002 is

more difficult to explain. The decrease in the first year coincided with an increase in

average size, as also seen at Île du Havre. A number of studies show that the growth

rate of S. droebachiensis increases with size (Swan, 1961; Thompson, 1984;

Raymond & Scheibling, 1987; Russell et al., 1998). Possibly, the small urchins at

Île à Firmin East benefited from the drop in large urchins that were either killed or

moved away due to ice scour. The reduced intraspecific competition and increased

food availability (with the new growth of kelp in the barrens) should have favored

increased growth and thus increases in the average size. However, the change during

42

the second year did not suggest a density‐dependent response to resources as a shift

to smaller urchins accompanied the decrease in numbers.

We documented striking differences between the two intensively studied sites,

the most striking being the dynamics of the kelp‐barrens interface and the general

decline in urchin abundance. Unfortunately, we did not expect such extreme

variation (e.g., the establishment of A. cribrosum at the Île du Havre site) and thus did

not sufficiently replicate sampling for making definitive statements regarding the

causes of these differences. Still, we postulate that A. cribrosum can play a pivotal role

in determining the dynamics of the kelp‐barrens interface by providing a

herbivore‐resistant fringe at the lower edge of the kelp bed. The distribution of

A. cribrosum in shallow waters may, in turn, be controlled by ice scour, which occurs

frequently during winter and early spring along rocky shores in the St. Lawrence

Estuary and the Gulf of St. Lawrence (Archambault & Bourget, 1983;

Bergeron & Bourget, 1984; Bourget et al., 1985), and along the coast of Newfoundland

(Hooper, 1981; Keats et al., 1985). We saw evidence of ice scour at Île à Firmin East

during the winter of 2000‐2001, as five of the six 10‐cm steel bolts set into the bedrock

as markers for the benchmark transects were found bent in June 2001. There was also

evidence of ice scour (although less severe) in the winter of 2001‐2002, as two

out of six bolts were bent (the damaged bolts found in June 2001 had been replaced).

These observations are consistent with those of Parks Canada personnel

(Benoît Roberge and Nancy Dénommée, personal communications) who recorded

a solid ice sheet between the islands for only two weeks in 2001 (20 February to

6 March), whereas there was a solid cover for two months in 2002 (18 January to

17 March) (there is less ice scouring when the cover is solid). In contrast, ice scour

appeared to have been less at Île du Havre, as only three out of six bolts were bent

during the winter of 2000‐2001 (all three in the shallower part of the kelp bed) and

only a single bolt during the 2001‐2002 winter. If the condition of the bolts did reflect

less ice scouring at Île du Havre, decreased disturbance may explain the greater

survival of A. cribrosum at this site. The drop in urchin density observed at

Île à Firmin East during winter of 2000‐2001 is harder to explain but may reflect

43

migration to greater depths due to winter conditions. Severe ice scouring appears to

have occurred at the Île à Firmin West site during the winter of 2001‐2002 as there

were numerous denuded areas in the kelp community at shallower depths. More

detailed observations are needed, however, to relate ice conditions to urchin

aggregation and dispersion below the kelp bed.

Our study is the first to examine factors affecting the temporal stability in the

lower limit of kelp beds in the northern Gulf of St. Lawrence. We provide evidence

that sea urchin grazing plays a pivotal role in determining the lower limit of kelp.

However, our results also show that the intensity of urchin‐kelp interactions is both

site and time specific. For example, the tendency of an increased rate of kelp bed

regression with increased urchin biomass was only seen at one of the intensively

studied site (Île à Firmin East). We propose that perturbations by abiotic factors

(e.g., ice scouring and water motion) triggering important but localized changes in

urchin abundance are of primary importance for controlling kelp beds in the northern

Gulf, rather than the large‐scale effects of biotic factors as observed in other regions

(e.g., predation and disease, Estes et al., 1978; Simenstad et al., 1978; Pearse & Hines,

1979; Duggins, 1980; Miller, 1985; Scheibling, 1986; Scheibling & Hennigar, 1997;

Scheibling et al., 1999). These perturbations would allow kelp beds to re‐establish

over the barrens during periods of reduced grazing. To increase our global

understanding of the dynamics of such subtidal macroalgal communities, additional

long‐term studies are needed to examine the temporal stability of algal assemblages

in the barrens zone, as well as to investigate the mechanisms by which a few fleshy

macroalgae (e.g., A. cribrosum and Desmarestia viridis) withstand intensive grazing by

urchins.

Chapitre III

Spatial and temporal stability in algal assemblages on urchin barrens in the northern Gulf of St. Lawrence

45

Résumé

L’algue brune Agarum cribrosum est la seule grande macrophyte pérenne charnue

communément retrouvée en zone dénudée dans le nord‐ouest de l’Atlantique. Nous

avons examiné la stabilité spatiale et temporelle des agrégats d’A. cribrosum ainsi que

leur impact sur le recrutement algal dans l’archipel de Mingan, dans le nord du golfe

du Saint‐Laurent. Les agrégats étaient très stables dans l’espace et dans le temps avec

que de faibles variations entre les sites étudiés. L’abondance (pourcentage de

recouvrement) d’A. cribrosum dans deux aires de 144 m2 s’est accrue de 6.5‐11.4 % en

2 ans et la plupart des changements d’abondance sont survenus en marge des

agrégats. L’aire de la surface de petits (<13 m2) agrégats d’A. cribrosum augmentait de

~1.8 % mois‐1 bien que des augmentations hivernales marquées (>95 %) sont

survenues, principalement dues à la fusion d’agrégats adjacents en de plus grands

agrégats. Les agrégats d’A. cribrosum favorisent le recrutement algal tel qu’indiqué

par l’abondance des juvéniles d’A. cribrosum et de l’algue rouge Ptilota serrata, plus de

cent fois supérieure à l’intérieur qu’à l’extérieur des agrégats. L’élimination

expérimentale de la canopée d’A. cribrosum (portions de 1 m2) en bordure d’agrégats

n’a eu aucun effet à long terme sur l’abondance d’A. cribrosum: en 14 mois, l’algue

avait à nouveau colonisé la plupart des aires initialement fauchées malgré le

recrutement de la laminaire Alaria esculenta. Les agrégats d’A. cribrosum semblent

donc résilients à de petites perturbations mais pourraient dépendre des brouteurs

pour supprimer les espèces compétitrices. À l’inverse des juvéniles d’A. cribrosum,

l’abondance de P. serrata a diminué rapidement après l’élimination de la canopée

pour n’être, à la fin de l’étude, qu’à 17 % de la valeur observée dans le traitement avec

canopée. Nos résultats démontrent que les agrégats d’A. cribrosum sont une

importante composante temporelle et spatiale en zone dénudée, en dépit du

broutage intensif typique de l’oursin. Le maintien et l’expansion des agrégats

d’A. cribrosum et de la flore associée semblent dépendre d’interactions positives

(facilitation) avec les adultes d’A. cribrosum, eux‐mêmes capables de repousser

l’oursin.

46

Abstract

The brown alga Agarum cribrosum is the only large, perennial, fleshy macrophyte

commonly found on urchin‐dominated barrens in the northwestern North Atlantic.

We examined the spatial and temporal stability of A. cribrosum stands and their

impact on algal recruitment in the Mingan Islands, northern Gulf of St. Lawrence.

The stands were highly stable in space and time, with only small variations between

the study sites. The abundance (percentage cover) of A. cribrosum in two 144‐m2 areas

increased by 6.5‐11.4 % over a 2‐year period, and most changes in abundance

occurred at the edge of the stands. The surface area of small (<13 m2) single stands of

A. cribrosum typically increased by ∼1.8 % month‐1, although marked increases (>95 %)

occurred during winter, largely because adjacent stands merged into larger, single

stands. Mature stands of A. cribrosum appear to enhance algal recruitment, as juvenile

A. cribrosum and the understory red alga Ptilota serrata were orders of magnitude

more abundant inside than outside the stands. The experimental removal of the

A. cribrosum canopy (1‐m2 portions) at the border of stands had no long‐term effect on

the abundance of A. cribrosum as within 14 months, the alga had recolonized most

of the cleared areas in spite of recruitment of the kelp Alaria esculenta. Thus,

A. cribrosum stands appear resilient to small‐scale perturbations but may require

grazers to suppress competitors. In contrast to juvenile A. cribrosum, the abundance

of P. serrata rapidly decreased after canopy removal and by the end of the study had

fallen to just 17 % of the value seen in control areas with an intact canopy. Our results

demonstrate that A. cribrosum stands are a temporally and spatially important

component of urchin barrens in spite of the heavy grazing that typically occurs there.

Maintenance and expansion of A. cribrosum stands and associated flora appear

to depend on positive interactions (facilitation) with self‐defended adults

of A. cribrosum.

47

Introduction

Herbivory and competition often reduce the recruitment and survival of marine algae

and both these types of negative interactions are thought to play key roles in

structuring marine communities (Randall, 1961; Lawrence, 1975; Lubchenco &

Gaines, 1981; Harrold & Reed, 1985; Carpenter, 1986; Lewis, 1986; Morrison, 1988).

Less attention has been paid to the potential positive effects of grazing‐resistant algae

on other algal species that are vulnerable to grazing (e.g., by interfering with the

foraging of herbivores and providing microhabitats of increased algal recruitment

and survival) (but see Hay, 1986; Pfister & Hay, 1988; Gagnon et al., 2003b) in spite

of the increasing recognition of the importance of positive interactions (facilitation) in

structuring communities (Bertness & Callaway, 1994; Callaway, 1995;

Bruno & Bertness, 2001).

This bias is evident in polar and temperate regions, where herbivory

(Pearse & Hines, 1979; Dayton et al., 1984; Keats et al., 1990; Sivertsen, 1997;

Scheibling et al., 1999; Gagnon et al., in press) and competition (Dayton et al., 1984;

Johnson & Mann, 1988; Paine, 1990; Reed, 1990; Leinaas & Christie, 1996) can strongly

affect the abundance and distribution of macrophytes, although abiotic factors such

as light availability, water motion, nutrients, sedimentation, salinity and ice scour are

clearly important as well (see review in Dayton, 1985b; but also Hooper, 1981;

Dean & Jacobsen, 1984; Gerard, 1984; Ebeling et al., 1985; Keats et al., 1985). Intensive

hebivory has been well documented in the subtidal zone in the northwestern North

Atlantic where the green sea urchin, Strongylocentrotus droebachiensis, severely limits

the distribution of kelp beds (Breen & Mann, 1976a; b; Chapman, 1981; Miller, 1985;

Scheibling, 1986; Scheibling et al., 1999; Gagnon et al., in press). Kelps are often

restricted to shallow water refuges but rapidly extend to deeper water when urchins

are naturally or experimentally removed (Himmelman et al., 1983; Scheibling, 1986;

Keats et al., 1990; Gagnon et al., in press).

As elsewhere, the intensive grazing of S. droebachiensis has led to the formation

of extensive urchin barrens in the Mingan Islands (northern Gulf of St. Lawrence,

48

Canada) that generally lack fleshy macrophytes and are instead dominated by

calcified red algae (Gagnon et al., in press). The perennial kelp, Agarum cribrosum, and

the annual filamentous alga Desmarestia viridis, are the only two brown algae

regularly found on these barrens. Agarum cribrosum has a large holdfast, a fairly rigid

stipe, and a large crinkled frond that becomes thicker and tougher with age. Urchins

have a low preference for this alga which is thought to be due to chemical deterrents

(e.g., phenolics, Vadas, 1968; 1977; Larson et al., 1980; Himmelman, 1984;

Himmelman & Nédélec, 1990). Agarum cribrosum is most common on topographical

high points and in areas with strong currents, which suggest that these situations

provide advantages. Although A. cribrosum is most often found in small stands

(1‐10 m2) on urchin barrens, it sometimes forms large stands (>500 m2) at the lower

edge of shallow beds of the kelp Alaria esculenta in exposed areas

(P. Gagnon, personal observation). Elsewhere, A. cribrosum has been shown to be

competitively inferior to other kelp species and is thought to require some level of

grazing to suppress competitors (Vadas, 1968).

Distinct invertebrate assemblages are associated with stands of A. cribrosum

(Bégin, 2002) with urchin densities much lower than in the surrounding barrens, and

A. cribrosum has been shown to restrict urchin movements (Gagnon et al., in press).

This reduction in herbivore density and activity could influence the flora associated

with stands of A. cribrosum, but there is no information on abundance or dynamics of

other algal species on urchin barrens. Moreover, we know little about the temporal or

spatial stability of this conspicuous and important component of the urchin barrens.

In this study, we examine the temporal variability in the abundance and distribution

of A. cribrosum and associated algae on urchin barrens in the Mingan Islands.

Specifically, we (1) quantify temporal changes in the abundance and distribution

of A. cribrosum over large areas, (2) examine temporal variability in the size of

A. cribrosum stands, (3) quantify the distribution of juvenile A. cribrosum,

Ptilota serrata and urchins in the vicinity of stands of A. cribrosum, and (4) assess the

impact of an A. cribrosum canopy on the recruitment and abundance of algae.

49


Our study was conducted from 1998 to 2002 (with observations during the summer

and fall) in the subtidal zone on the eastern side of Île aux Goélands, and on the

western side of Petite île au Marteau, in the Mingan Islands, northern Gulf of

St. Lawrence (Fig. 2.2). At each site, the bottom is a gently‐sloped bedrock platform

which is covered by dense kelp bed, primarily Alaria esculenta, to a depth of 3‐5 m.

Below this shallow kelp bed an urchin barrens begins and extends a considerable

distance (~200 m from the shore at Île aux Goélands and 150 m at Petite île au

Marteau) to a depth of ~10 m, where a vertical wall begins. The only fleshy

macrophytes commonly found on the barrens at the two sites were the brown algae

Agarum cribrosum and Desmarestia viridis although small patches of the red

alga Ptilota serrata were also present, especially at greater depths.

Temporal variability in abundance and distribution of A. cribrosum

To evaluate the variability in the abundance and distribution of Agarum cribrosum, we

monitored changes in 12 x 12‐m (144 m2) areas in the barrens at the Île aux Goélands

and Petite île au Marteau sites. These sites were selected after exploratory surveys

based on the similarity in stand configuration, i.e., distinct stands of A. cribrosum

separated by large areas without fleshy algae. Although the abundance

of A. cribrosum was somewhat higher at these sites than on most urchin barrens in this

region, these sites were selected because the clear boundaries of the stands increased

our ability to detect changes in the abundance and distribution of A. cribrosum, and

the proximity of the sites to our laboratory facilitated sampling.

At each site, we permanently marked the corners of the sampling area with

bolts set into the bedrock. We subdivided each area into an 8 x 8 sampling grid of

64 1.5 x 1.5‐m contiguous plots using the following procedure. We first delineated

two facing sides of the area with 12‐m benchmark lines tied to bolts and marked at

1.5‐m intervals. A 1.5 x 12‐m section was then defined by attaching two 12‐m transect

lines to the benchmark lines and then subdivided the section into 8 plots using

a 1.5 x 1.5‐m quadrat positioned using the marks at 1.5‐m intervals on the transect

50

lines. Transect lines were then shifted to the adjacent 1.5 x 12‐m section and the

sampling repeated until the entire area was sampled. Within each plot, the same

diver estimated the percent cover of A. cribrosum at 5 times over a 2‐year period

(June 2000, August 2000, June 2001, August 2001, and June 2002). Benchmarks and

transect lines were removed between sampling periods.

Temporal variability in the size of A. cribrosum stands

To evaluate changes in the size of stands of Agarum cribrosum over time, we

monitored the surface area of 12 haphazardly‐selected stands of adult A. cribrosum

near the permanent 12 x 12‐m area at the Île aux Goélands site. For each stand, we

marked its position with a bolt set into the substratum at the edge of the stand, and

then recorded (to the nearest 1 cm) (1) the longest distance across the stand (A), and

(2) the longest distance running perpendicular to the first axis (B) based on the area

covered by the canopy. Given that most of the stands were ellipsoidal in shape, we

approximated the surface area (S) of each stand using the equation S = π (A/2) (B/2).

Sampling was done 5 times over a 2‐year period (July 1999, May 2000, August 2000,

June 2001 and July 2001). Initial size of stands ranged from 1.6 to 12.8 m2.

Distribution of algae and urchins near A. cribrosum

To determine whether stands of adult Agarum cribrosum are associated with fewer

urchins and increased algal recruitment, we compared the abundance of juvenile

A. cribrosum (i.e., <10 cm long), Ptilota serrata and urchins at the center of the stand

with that at different distances from the edge of stands. Six haphazardly‐chosen

A. cribrosum stands (average width of 2.4 m [SD = 0.6] and separated by at least 4 m)

were sampled on the eastern side of Île aux Goélands in July 2000. For each stand, we

counted the numbers of juvenile A. cribrosum and urchins and estimated the percent

cover of P. serrata in 25 x 25‐cm quadrats placed east and west of the point which

we considered was the center of the stand, and also at 0, 50 and 100 cm from the

eastern and western edges of the stand. The data from the quadrats in two directions

were pooled in the analysis.

51

Effect of A. cribrosum canopy on recruitment and abundance of algae

To assess the impact of Agarum cribrosum on the recruitment and abundance of algae

on urchin barrens, we monitored changes in the density of juvenile A. cribrosum and

percent cover of Ptilota serrata in the absence and presence of a canopy of adult

A. cribrosum over 14 months. The two treatments consisted of 1‐m2 areas where

A. cribrosum was experimentally removed (cut off at the holdfast) and controls

where the A. cribrosum were left intact. For this experiment, we used two large

(>50 m2) oblong stands of A. cribrosum that ran parallel to the shore at 3‐4 m in depth

at Île aux Goélands. In August 1998 we selected 16 1‐m2 plots, spaced at 1‐2 m

intervals along the lower edges of the stands, and marked each with a bolt in the

center. The plots were grouped into 8 blocks of two adjacent 1‐m2 areas, and one of

the two was randomly chosen for the canopy removal treatment. Stands were then

sampled 5 times: August 1998 (1 d after canopy removal), October 1998, June 1999,

August 1999 and October 1999. At each time, we estimated the number of juvenile

A. cribrosum and the percent cover of P. serrata in each plot using two quadrats

(25 x 25 cm) placed on the north and south sides of the center bolt. Following a heavy

recruitment of the kelps A. cribrosum and Alaria esculenta during the winter of

1998‐1999, we were not able to separate juveniles and adults in 1999 due to logistic

constraints. We averaged the data from the two quadrats in each area before

comparing species abundance among treatments.


We used a repeated measures ANOVA (Hand & Taylor, 1987; Crowder & Hand,

1990) to examine temporal changes in the percent cover of A. cribrosum in the

8 x 8 observation plots on the barrens with the factors Site (Île aux Goélands

and Petite île au Marteau) and Time (5 sampling dates), and with one nested element

(Plots within Site). To take into account the spatial dependency among the 64 plots

at each site, we modeled the covariance structure with an exponential correlation

structure (Proc Mixed, type = SP(exp)(longitude latitude), SAS Institute Inc., 1999).

This structure assumed that the correlation between plots decreased with the distance

52

separating them. Note that the information obtained for the 64 dependent plots was

the same as for 6 independent plots (Lettenmaier, 1976). The temporal dependency

was modeled with a first order autoregressive structure (Proc Mixed,

type = SP(pow)(month), SAS Institute Inc., 1999) common to both sites. Since no

transformation corrected the lack of normality in the data on some observation dates,

the ANOVA was also run with the rank transformed data. As both analyses gave

similar results, we presented the results from the analysis of the raw data

as suggested by Conover (1980).

We analyzed temporal changes in the size of stands of A. cribrosum using

a repeated measures ANOVA with the factor Time (5 sampling periods). Because four

of the pairs of stands had merged together by June 2001 (and remained merged until

the end of our study), the surface area of four stands was given a value of nul (0) in

the analysis in June and July 2001. The surface area of each of the pairs of stands that

had merged was attributed to the larger stand at the beginning of the study

(this minimized the increase in size). Since variances fluctuated over time,

we modeled the covariance structure with a specific heterogeneous compound

symmetry structure (Proc Mixed, type = CSH, SAS Institute Inc., 1999). We applied

the analysis to the raw data as the data for each sampling date were normally

distributed.

To evaluate the effect of the distance from stands of Agarum cribrosum on the

abundance of juvenile A. cribrosum, Ptilota serrata and urchins, we ran one‐way

ANOVAs (Zar, 1999) with the factor Distance (center of the stand, and 0, 50, and

100 cm from the edge of the stand). The analyses were applied to the raw data for

urchins whereas numbers of juvenile A. cribrosum were square‐root transformed to

correct for heteroscedasticity. Since no transformations corrected for the

heteroscedasticity detected in the raw data on the percent cover of P. serrata, the

ANOVA was also applied to the rank transformed data. As both analyses gave

similar results, we presented the raw data (Conover, 1980).

53

To examine temporal changes in the density of A. cribrosum juveniles and

P. serrata in the absence and presence of a canopy of A. cribrosum, we conducted

repeated measures ANOVAs with the factors Treatment (with/without canopy),

Stand (2 stands of A. cribrosum), Block (8 blocks of two treatments), and Time

(August and October 1998, and June, August, and October 1999), and with one nested

element (Block within Stand). Since variances were stable over time within the stands,

and the correlation in the data over time was similar between the treatments, we used

a pooled covariance structure in the ANOVA (Proc Mixed, type = CS, SAS Institute

Inc., 1999). We applied the analysis to the raw numbers on A. cribrosum as the data for

each time were normally distributed, whereas the data for the percent cover of

P. serrata were square‐root transformed to obtain normality for each observation date.



distribution of the residuals and by applying Levene test (Snedecor & Cochran, 1989).


multiple comparisons tests (LS means, SAS Institute Inc., 1999). A significance level

of 0.05 was used for all statistical tests.

Results

Temporal variability in abundance and distribution of A. cribrosum

The 2‐year monitoring of the 12 x 12‐m observation areas on the barrens indicated

only small overall increases in the abundance of Agarum cribrosum at both sites

(Table 3.1, Fig. 3.1). At Île aux Goélands, the abundance of A. cribrosum decreased

slightly during the first two months but increased steadily (~0.7 % month‐1) after

August 2000 and was 11.4 % greater at the end than at the beginning of the study

(LS means, P < 0.0001). At Petite île au Marteau, more substantial increases and

decreases occurred during the study, including a ~25 % increase between August

2000 and June 2001, but the percentage cover of A. cribrosum at the end of the study

was 6.5 % greater than at the beginning.

54

Table 3.1. Summary of the repeated measures ANOVA (mixed linear model) showing the effect of Site (Île aux Goélands and Petite île au Marteau) and Time (June and August 2000, June and August 2001, and June 2002) on the percent cover of A. cribrosum in 12 x 12‐m areas of urchin barrens (as averaged from 64 plots of 1.5 x 1.5 m at each site).

Source of variation

(fixed effects)

df

(numerator)

df

(denominator)

F‐value

P

Site 1 126 0.33 0.56

Time 4 504 42.26 <0.0001

Site x Time 4 504 12.37 <0.0001

55

0

20

40

60

80

100

Jun 2000 Aug 2000

Petite île au Marteau

Île aux Goélands

Jun 2001 Aug 2001 Jun 2002

aba

bc

d

ABBC

DC

******

** *

Fig. 3.1. Changes, over 2 years, in the mean percent cover (± SE) of kelp Agarum cribrosum on urchin barrens at Île aux Goélands and Petite île au Marteau. Values not sharing the same letter (small letters for Petite île au Marteau, capitals for Île aux Goélands) are different (LS means, P < 0.05). Asterisks are from statistical comparisons between islands in the increase or decrease in percent cover during 4 successive sampling periods (Group contrasts, *P < 0.05, **P < 0.01, ***P < 0.001).

56

The distribution of A. cribrosum appeared more stable at Île aux Goélands

(Fig. 3.2). The general gradual increase in A. cribrosum was due to colonization along

the edges of stands which were around one large area of barrens (in the upper

right corner on Fig. 3.2). Agarum cribrosum was more patchily distributed at Petite île

au Marteau and the observed fluctuations in abundance occurred across the grid.

Temporal variability in the size of A. cribrosum stands

During the first year of the 2‐year monitoring of the size of stands of

Agarum cribrosum, there was no significant change in area (LS means, P = 0.22).

In contrast, stand area increased greatly (13 % month‐1) between August 2000 and

July 2001 (Fig. 3.3; LS means, P = 0.019) due primarily to the merging of stands after

incremental increases at the margins of stands (all the stands merged with other

stands; four of the mergers occurring between stands we were monitoring whereas

the remaining four stands merged with stands that we were not monitoring).

The 17.1 % increase in size of stands in the month following this massive increase

(June to July 2001) was not statistically significant.

Distribution of algae and urchins near A. cribrosum

An inverse pattern of abundance of algae and urchins occurred inside and around

stands of adult Agarum cribrosum (Fig. 3.4). Numbers of juvenile A. cribrosum and

percentage cover of P. serrata were orders of magnitude greater inside than outside

the stands, whereas urchin densities were ≥5 times less inside than outside the stands.

No differences in the abundance of urchins and algae were seen at different distances

sampled from the edge of the stands (sampling at 0, 50 and 100 cm, Fig. 3.4),

consistent with the idea that A. cribrosum stands provided areas with decreased

urchins and favored algal recruitment or survival.

Effect of A. cribrosum canopy on recruitment and abundance of algae

The 14‐month monitoring of changes in the numbers of juvenile Agarum cribrosum

57

Petite île au MarteauÎle aux Goélands

Jun 2000

Aug 2000

Jun 2001

Aug 2001

Jun 2002

0

10

20

30

40

50

60

70

80

90

100

Percent cover (%)

Fig. 3.2. Changes, over 2 years, in the distribution and abundance of kelp Agarum cribrosum in a 12 x 12‐m zone of urchin barrens at Île aux Goélands and Petite île au Marteau. Dashed lines represent the position of transect lines used to divide each zone into a sampling grid of 64 1.5 x 1.5‐m plots.

58

0

5

10

15

20

Jul 1999 May 2000 Aug 2000 Jun 2001 Jul 2001

a

a

bbb

Fig. 3.3. Changes, over 2 years, in the mean surface area (+ SE) of stands of adult kelp Agarum cribrosum. The arrow indicates the time at which we first observed mergers between the stands. Values not sharing the same letter are different (LS means, P < 0.05).

59

Centerof stand

0 50 100Distance (cm)

0

50

100

150

200

250

0

10

20

30

40

50

0

100

200

300

400

500 S. droebachiensis

Juvenile A. cribrosum

P. serrata

b

a a

a

a

b b b

a

b b b

Edge

OutsideInside

Fig. 3.4. Mean density and percent cover (+ SE) of juvenile kelp Agarum cribrosum, the red alga Ptilota serrata, and the green sea urchin, Strongylocentrotus droebachiensis, at different distances from the center of stands of adult A. cribrosum. For juvenile A. cribrosum, the data have been back‐transformed from the transformed data used in the analysis. Values not sharing the same letter are different (LS means, P < 0.05).

60

and percentage cover of Ptilota serrata in the absence and presence of a canopy

of A. cribrosum indicated that the canopy had little or no effect on the recruitment of

A. cribrosum but had a large effect on the abundance of P. serrata (Table 3.2, Fig. 3.5).

The density of juvenile A. cribrosum did not differ between treatments except in

June 1999 when densities were almost double in the presence of a canopy.

Unfortunately, as we did not distinguish between adult and juveniles during this and

subsequent samplings, part of this difference is due to the inclusion of the adult

plants that formed the canopy itself. However, given that densities of adults are

typically near 50 m‐2 and the generally high recruitment of both A. cribrosum and

Alaria esculenta (see below) that occurred between October 1998 and June 1999, we

attribute this difference primarily to higher recruitment in treatments with an adult

canopy. Regardless, this difference did not persist and the two treatments were

statistically indistinguishable by the end of the season.

In contrast, differences in the abundance of P. serrata were apparent from the

beginning and persisted until the end. One day after the removal of the canopy, the

mean percent cover of P. serrata fell to less than 50 % of that seen in areas with an

A. cribrosum canopy although this difference was not significant (LS means, P = 0.10).

In the treatment with a canopy, the cover of P. serrata generally increased during the

study and by October 1999 was 53 % greater than at the beginning of the study

(Fig. 3.5). Although there was not a statistical difference between the first and last

date of the study, cover was significantly greater on the penultimate sampling date

(LS means, P = 0.0054). In contrast, in the areas with the canopy removed the cover

of P. serrata remained low (<5.3 %) throughout the study and was only 17 % of that

in control areas by the end of the study (LS means, P = 0.0003; Fig. 3.5).

In both areas with and without a canopy, the percent cover of other

non‐calcified algae (e.g., ulvoid algae, diatoms, and the red algae Phycodrys rubens

and Porphyra sp.) was always low (<5 %). An exception was a substantial recruitment

of the kelp Alaria esculenta between October 1998 and June 1999 that appeared to be

61

Table 3.2. Summary of repeated measures ANOVA (mixed linear model) showing the effect of Stand (two stands), Treatment (with or without a canopy of Agarum cribrosum), and Time (August and October 1998, and June, August, and October 1999) on the mean density of A. cribrosum and percent cover of Ptilota serrata in the A. cribrosum canopy‐removal experiment. Data on P. serrata were square‐root transformed.

Source of variation (fixed effects)

df (numerator)

df (denominator)

F‐value

P

Agarum cribrosum Stand 1 12 0.97 0.34 Treatment 1 12 4.88 0.047 Stand x Treatment 1 12 0.89 0.36 Time 4 47 23.12 <0.0001 Stand x Time 4 47 5.24 0.0014 Treatment x Time 4 47 1.84 0.14 Stand x Treatment x Time 4 47 0.14 0.96 Ptilota serrata Stand 1 12 2.61 0.13 Treatment 1 12 13.85 0.0029 Stand x Treatment 1 12 0.17 0.69 Time 4 47 2.08 0.10 Stand x Time 4 47 1.60 0.19 Treatment x Time 4 47 4.18 0.0056 Stand x Treatment x Time 4 47 0.10 0.98

62

0

50

100

150

200

250

300

350 A. cribrosum

With A. cribrosum

Without A. cribrosum

0

5

10

15

20

25

30

35

40 P. serrata

Aug 1998 Oct 1999Aug 1999Jun 1999Oct 1998

cc

a a

b

cc

b

a

bc

BC

BC

B

A

AB

C D CDD D

Fig. 3.5. Changes, over 14 months, in the mean density (+ SE) of kelp Agarum cribrosum and percent cover of the red alga Ptilota serrata in the absence and presence of a canopy of A. cribrosum. The arrow indicates the time from which the sampling was conducted without discrimination between juvenile and adult stages of A. cribrosum. Values not sharing the same letter (small letters for A. cribrosum, capitals for P. serrata) are different (LS means, P < 0.05).

63

greater in the presence (93.0 m‐2, SD = 69.4) than absence (65.1 m‐2, SD = 24.2) of

an A. cribrosum canopy. This difference was not significant (LS means, P = 0.39),

but A. esculenta recruits completely disappeared from areas with a canopy by

August 1999 whereas they were still common in treatments where the canopy was

removed (38.0 m‐2, SD = 18.3). At the end of the study (October 1999), A. esculenta was

absent in both treatments.

Discussion

Agarum cribrosum, the only large, perennial, fleshy macrophyte and the only kelp

species commonly found on urchin‐dominated barrens in coastal areas of the

northwestern North Atlantic (Tremblay & Chapman, 1980; Keats et al., 1982;

Himmelman & Lavergne, 1985; Novaczek & McLachlan, 1989), appears to play a

considerable role in determining the structure of the barrens community. A diverse

and distinct community of invertebrates is associated with A. cribrosum, and its

presence can increase the recruitment and growth of a number of invertebrate species

including the mussel Mytilus edulis and the polychaete Spirorbis borealis (Bégin, 2002).

In contrast, the effect of A. cribrosum on the subtidal flora has not been investigated,

and the spatial and temporal stability of A. cribrosum stands is unknown over any

spatial or temporal scale. Here, we documented a relatively high degree of temporal

stability in the abundance and distribution of A. cribrosum as well as a tight

association between A. cribrosum and other fleshy macroalgae in the Mingan Islands.

Our 2‐year monitoring of A. cribrosum in the 12 x 12‐m observation areas at

Île aux Goélands and Petite île au Marteau showed that A. cribrosum can be

a dominant feature of urchin barrens. It had a percent cover of >45 % throughout the

study and was capable of increasing in abundance in spite of the intensive grazing by

urchins (Gagnon et al., in press). The increases were due to gradual increases in size

of currently existing stands, rather than to the formation of new stands, with

expansion or recession occurring along the edge of stands and no change in the

central areas. Our observations of individual stands in an adjacent area at

Île aux Goélands are consistent with these larger scale trends as the area covered by

64

A. cribrosum increased both at stand margins and by the merging of stands. Thus,

established stands appear to be persistent (sensu Dayton et al., 1984) features of the

landscape (with a tendency towards a general increase in size through time) although

they have the potential for substantial short‐term decreases and increases in size.

Much of our knowledge about the ecology of A. cribrosum comes from the

pioneering work of Vadas (1968) in the San Juan Islands, northeastern Pacific.

He found that (1) the growth of A. cribrosum is greatest during late fall, winter, and

early spring (when water temperatures are at their lowest), (2) A. cribrosum can

become established during the fall and winter seasons, when the majority of potential

competitors (other kelp species) are unable to develop, and (3) A. cribrosum is

competitively inferior to other kelp species and can only become established when

urchins (S. droebachiensis and Strongylocentrotus franciscanus) are present. Our

experimental removal of the A. cribrosum canopy are partially consistent with these

observations. We too observed the greatest recruitment during the winter months,

but in contrast to previous work, we observed that A. cribrosum was able to recover

in spite of colonization by other kelp species (e.g., Alaria esculenta). The explanation

for this resilience is not totally clear. When A. esculenta first appeared in June 1999, the

density of A. esculenta recruits was less than half that of A. cribrosum, giving

A. cribrosum a numerical advantage if competition was important. A more likely

explanation is selective herbivory by the green sea urchin, which consumes

A. cribrosum at a much lower rate than preferred algae, such as species of Alaria and

Laminaria (Vadas, 1977; Larson et al., 1980; Himmelman, 1984; Himmelman

& Nédélec, 1990). Consistent with this idea are our observations from a concurrent

study in which A. cribrosum replaced A. esculenta (Gagnon et al., in press). This shift in

algal dominance occurred at the lower edge of an A. esculenta bed where urchins had

formed a grazing front. Over two years, the lower portion of a large

A. esculenta‐dominated zone was gradually replaced by A. cribrosum. In the present

study, the gradual decrease in the density of A. cribrosum from June to October 1999

to the levels at the beginning of the study may have been due to loss to grazers.

Alternatively, it was possibly also due to the loss of individuals from intraspecific

65

competition for resources (e.g., space, light and nutrients). Overall, our observations

indicate that A. cribrosum is resilient to punctual, small‐scale (~1 m) disturbances,

which should permit A. cribrosum to recover from disturbances by ice scour,

a common phenomenon in Newfoundland and the Gulf of St. Lawrence (Hooper,

1981; Bergeron & Bourget, 1984; Keats et al., 1985; Gagnon et al., in press).

The spatial and temporal stability of A. cribrosum stands should favor

associations with algae or algal stages which are more vulnerable to grazing or to

dislodgment by currents and waves. Our observation that A. cribrosum juveniles and

the red alga Ptilota serrata recruited within the stands of adult A. cribrosum, where

urchins densities are reduced, suggests that A. cribrosum stands provide understory

algae with protection from urchin grazing. However, our experiments only partially

support this idea. Recruitment and recovery of A. cribrosum did not require the

presence of an adult canopy although it appears that recruit densities of both

A. cribrosum and A. esculenta were higher in the presence of the canopy during

the initial phases of colonization (i.e., during the winter period). Alternatively, the

association of recruitment of A. cribrosum with adult stands could be more linked to

a higher density of gametophytes under the canopy. The removal of the canopy

would then not have had an immediate effect on recruitment as there may have been

a bank of gametophytes already present when the experiment was initiated.

In contrast, there was a clear positive effect of A. cribrosum on P. serrata

as removal of the canopy lead to a rapid and persistent decrease in its abundance.

More detailed investigations are needed to determine the exact mechanism by which

A. cribrosum might enhance the recruitment of P. serrata. Reduced grazing is certainly

a strong possibility, and previous studies have shown that mechanical abrasion from

wave‐induced movement of algal fronds repulses urchins (Konar, 2000;

Gagnon et al., 2003b). Whether the stiffer and more upright fronds of A. cribrosum

provide the same type of protection remains to be seen. Alternatively, it is possible

that P. serrata is more abundant under A. cribrosum because the canopy prevents

it from being broken and dislodged by waves and currents. Ptilota serrata is often

66

loosely attached (P. Gagnon, personal observation), and algal canopies have been

shown to reduce flow (Eckman et al., 1989; Ackerman & Okubo, 1993). This

mechanism could explain the decrease in the abundance of P. serrata immediately

after the removal of the A. cribrosum canopy. The positive effects of stands of

A. cribrosum on the recruitment of algae on urchin barrens did not apply

to Desmarestia viridis, the other common brown alga of this habitat. This annual

macroalga was not associated with stands of A. cribrosum and appears to colonize the

barrens by recruiting in the winter when urchin grazing activity is low and then

relying on mechanical repulsion of urchins once it reaches a larger size (Chapter 5).

A critical remaining question is how stands of A. cribrosum are initially formed.

We have previously shown that juvenile A. cribrosum are avidly consumed by

urchins, but protected to a degree when associated with D. viridis (Gagnon et al.,

2003b). However, as D. viridis is an annual, this protection is ephemeral. It remains

unknown how juvenile stages can survive for longer periods outside stands of adult

plants. Perhaps an exceptional year of massive recruitment is required, but only

longer term and larger scale studies are likely to be able to answer this question.

Our study is the first to examine the temporal changes in algal assemblages on

urchin barrens in the northwestern North Atlantic. We have shown that the spatial

and temporal stability of stands of A. cribrosum can be high, contrary to the low

stability of kelps which are readily grazed by the green sea urchin (Breen & Mann,

1976b; Scheibling et al., 1999; Gagnon et al., in press). This stability may be due, in

part, to the enhanced recruitment of juveniles in adult stands of A. cribrosum. These

stands also increase algal diversity by increasing the abundance of P. serrata. As

demonstrated for D. viridis (Gagnon et al., 2003b), this situation represents a case of

facilitative interaction (Connell & Slatyer, 1977; Callaway, 1995; Bruno & Bertness,

2001) that permits algal recruitment and survival on urchin barrens. More studies are

needed to determine how A. cribrosum and D. viridis withstand urchin grazing during

juvenile stages and to estimate the relative contribution of adult A. cribrosum and

D. viridis in maintaining algal assemblages in this intensively‐grazed environment.

Chapitre IV

Algal colonization in urchin barrens: defense by association during recruitment of the brown alga Agarum cribrosum

68

Résumé

Nous avons examiné le processus par lequel les juvéniles de la laminaire

Agarum cribrosum échappent au broutage de l’oursin vert, Strongylocentrotus

droebachiensis, dans la zone rocheuse infralittorale dénudée de l’archipel de Mingan,

dans le nord du golfe du Saint‐Laurent. Le recrutement le plus élevé de juvéniles

d’A. cribrosum avait lieu sous la canopée de la grande phaéophycée filamenteuse

Desmarestia viridis, où les densités d’oursins étaient considérablement réduites

comparativement à l’aire environnante. Ces patrons de distribution semblaient liés au

mouvement de balayage de D. viridis, bien que les courants peuvent modifier le

mouvement de va‐et‐vient de l’algue en poussant la canopée dans une direction

spécifique, permettant alors aux oursins d’envahir les aires non balayées. La densité

de juvéniles d’A. cribrosum sous les plants de D. viridis augmentait avec la taille du

plant et avec l’accroissement de la proximité au crampon. Les juvéniles d’A. cribrosum

vivant sous une canopée de D. viridis affichaient un taux de croissance légèrement

inférieur aux juvéniles dépourvus d’une canopée mais cette perte de croissance était

nettement compensée par le gain de protection contre le broutage de l’oursin.

La période pendant laquelle D. viridis fournit une protection aux juvéniles est de

l’ordre de quelques mois car D. viridis est une espèce annuelle qui disparaît en début

d’automne. Cette défense des juvéniles d’A. cribrosum par association avec D. viridis

est possiblement une étape dans la succession menant à la formation de plaques

matures d’A. cribrosum.

69

Abstract

We investigated the process whereby juveniles of the kelp Agarum cribrosum escape

grazing by the green sea urchin, Strongylocentrotus droebachiensis, on urchin barrens in

the rocky subtidal zone in the Mingan Islands, northern Gulf of St. Lawrence. The

highest recruitment of juvenile A. cribrosum occurred under the canopy of the large

filamentous phaeophyte Desmarestia viridis, where urchin densities were markedly

reduced, compared to the surrounding area. These patterns of distribution appeared

to be related to the wave‐induced sweeping motion of D. viridis, although currents

may modify the back and forth motion of the alga by pushing the canopy towards a

specific direction, thereby allowing urchins to invade the non‐swept areas.

The density of juveniles of A. cribrosum under D. viridis plants increased with plant

size and increasing proximity to the holdfast. Living under D. viridis slightly reduced

the growth rate of the A. cribrosum juveniles, but this loss in growth was clearly

outweighed by the gain in protection from sea urchin grazing. The time scale over

which D. viridis provides protection to the juveniles is in the order of months, as

D. viridis is an annual alga that disappears in early autumn. This defensive

association of juvenile A. cribrosum with D. viridis is possibly a successional step

leading to the formation of mature stands of A. cribrosum.

70

Introduction

Plants exhibit two general resistance strategies against herbivory: tolerance and

avoidance (Rosenthal & Kotanen, 1994). Whereas mature plants can use both

strategies, immature plants cannot usually tolerate a substantial loss of tissue to

herbivores and therefore must avoid being eaten, either by producing defenses or by

escaping in space or time (Rosenthal & Kotanen, 1994). A plant defense is any

attribute that negatively influences the feeding behavior or abundance of potential

herbivores (OʹDowd & Williamson, 1979). The production of chemical deterrents and

of grazing resistant structures are defenses used to reduce the foraging success

of herbivores (Paul & Hay, 1986; Hay & Fenical, 1988; Van Alstyne, 1988; Lucas et al.,

2000; Schnitzler et al., 2001). Spatial and temporal escapes are also a well known

phenomenon. One common example is defense by association, which occurs when a

palatable species exploits micro‐sites with lowered herbivory created by the presence

of an unpalatable species (Tahvanainen & Root, 1972; Atsatt & OʹDowd, 1976;

OʹDowd & Williamson, 1979; Hay, 1986; Brown & Ewel, 1987; Pfister & Hay, 1988).

Although this strategy might lead to competitive costs (e.g., lowered growth rates) for

the protected species, the prevention of local extinction is an obvious advantage

(Hay, 1986).

Sea urchins are often the dominant herbivore in marine rocky habitats and

may strongly influence the structure of algal assemblages by reducing the growth and

survival of algae (see review by Lawrence, 1975; but also Paine & Vadas, 1969;

Dean et al., 1984; Bulleri et al., 1999; Rose et al., 1999). In eastern Canada, intensive

grazing by the green sea urchin, Strongylocentrotus droebachiensis, often leads to the

formation of subtidal barrens with sparse algal cover (Breen & Mann, 1976b;

Scheibling et al., 1999). Such barrens are extensive in shallow areas of the Mingan

Islands, northern Gulf of St. Lawrence, where urchin densities often attain

300 individuals m‐2 (Himmelman & Nédélec, 1990), and where only a few algal

species can withstand the intensive grazing (e.g., the red alga Ptilota serrata and

crustose corallines Lithothamnion sp. and Clathromorphum sp.).

71

Of the larger brown algae, only two species, Agarum cribrosum and

Desmarestia viridis, commonly occur on urchin barrens. Agarum cribrosum is a

perennial alga which has a large holdfast, a semi‐rigid stipe, and a large crinkled

frond which becomes thicker and tougher with age (Vadas, 1968). Its low palatability

to urchins is thought to be due to chemical deterrents in the tissues (e.g., phenolics,

Vadas, 1977; Larson et al., 1980; Himmelman & Nédélec, 1990). Vadas (1968) also

indicates that A. cribrosum is most often found in areas with strong currents. In the

Mingan Islands, we have noted that A. cribrosum is usually found in small patches

(1‐10 m2) within urchin barrens and is more common on topographical high points. In

contrast, D. viridis is an annual alga with a small holdfast, a highly flexible stipe and a

large, profusely‐branched frond which sweeps back and forth over the bottom with

even slight wave action. The sweeping motion appears to repel urchins as they

rapidly aggregate on the alga during calm conditions (Himmelman & Steele, 1971;

Himmelman, 1984; Konar, 2000; Chapter 5). Many species of Desmarestia, including

D. viridis, produce sulfuric acid (Eppley & Bovell, 1958; McClintock et al., 1982;

Thompson, 1988), and numerous studies infer that this might also deter grazers

(Anderson & Velimirov, 1982; Dayton, 1985a; Himmelman & Nédélec, 1990; Lobban

& Harrison, 1994). However, the only experimental evidence supporting this

hypothesis is a recent study on Desmarestia munda (Pelletreau & Muller‐Parker, 2002).

The exact mechanism by which D. viridis repels urchins remains unclear.

Early stages of algae are often particularly vulnerable to herbivores, because of

their small size and more delicate structure (Van Alstyne et al., 1999). We have a poor

understanding of the factors permitting the survival of juvenile A. cribrosum and

D. viridis. The rapid growth of juvenile D. viridis in late winter and early spring, when

feeding rates of the urchins are low (Larson et al., 1980; Himmelman, 1984), may

permit it to reach the size where the sweeping action (or acid content) provides

resistance to grazing (thus an escape in size, sensu Paine, 1976; Lubchenco & Gaines,

1981). On the other hand, juvenile A. cribrosum (defined here as individuals

measuring <10 cm in length) appear to be unable to survive on their own as they

rarely occur in open areas on urchin barrens. However, we have often observed

72

juvenile A. cribrosum under stands of adult A. cribrosum and also under canopies of

D. viridis, which suggests that recruitment under these algae may represent a type of

defense by association. Such a relationship could provide a mechanism whereby

patches of A. cribrosum can become established.

In the present study we investigate this potential associational plant defense to

determine how A. cribrosum might colonize urchin barrens. Specifically, we

(1) quantify the distribution of the urchins and juvenile A. cribrosum in the vicinity of

D. viridis, (2) determine whether the size of the D. viridis plant influences these

distributions, (3) measure the degree of vulnerability of juvenile A. cribrosum to

urchin grazing, (4) examine whether there is a competitive cost for juvenile

A. cribrosum in being associated with D. viridis, and (5) assess the time scale involved

in these interactions.


Study sites

Our study was conducted during the summer and fall of 1997, 1998 and 1999 in the

subtidal zone of Île aux Goélands, a small (<1 km2) exposed island in the Mingan

Islands, northern Gulf of St. Lawrence (Fig. 2.2). On the bedrock platforms on both

eastern and western sides of the island, dense kelp beds (primarily Alaria esculenta)

occur to a depth of 3 m, and then urchin barrens dominate gently sloping platforms

that extend to ∼200 m from shore. These platforms end at ∼10 m in depth, where

vertical walls begin. On the western platforms, Desmarestia viridis is abundant and

Agarum cribrosum scarce, and on the eastern platforms, D. viridis is less abundant and

A. cribrosum more abundant. Urchin predators, such as lobsters (Homarus americanus)

and wolffish (Anarhichas lupus) have never been observed on the platforms, and the

other conspicuous herbivores are small limpets and chitons which are micro‐grazers

and generally less abundant than urchins (Gaymer et al., 2001; P. Gagnon, personal

observation).

73

Distribution of juvenile A. cribrosum and urchins near D. viridis

We first examined the impact of the sweeping action of Desmarestia viridis in July and

August of 1997 by comparing the abundance of urchins and juvenile

Agarum cribrosum (<10 cm in length) inside and outside the area swept by individual

D. viridis plants. This was done in the vicinity of nine haphazardly‐chosen,

averaged‐sized (50‐60 cm in length) D. viridis plants. Each plant was at least 2 m from

other plants (holdfast to holdfast). For each plant we counted the numbers of urchins

and A. cribrosum juveniles in two quadrats (25 x 25 cm) placed next to the holdfast

(on the north and south sides, respectively) and in two quadrats placed 25 cm outside

of the area swept by the plant (again on the north and south sides). Six of the plants

were on the eastern side of Île aux Goélands (three on each of two sampling dates)

and the other three on the western side (one sampling date).

Current and wave conditions cause the sweeping action of D. viridis and may

in turn affect the distribution of urchins and A. cribrosum juveniles. We therefore

measured current conditions on the two sides of Île aux Goélands using Aanderaa

RCM4 current meters. The current meters were anchored at 1 m above the bottom

(4.4 m in depth on the eastern side and 5.3 m on the western side) with rigid supports

and recorded at 15‐min intervals from 26 June to 19 August of 1999. A plastic conical

fence below the current meters prevented crabs and urchins from climbing onto the

apparatus. Wave action was not quantified but is generally higher on the western

side of the island due to the predominant winds from the west (meteorological

airport services from Havre‐Saint‐Pierre Airport), except during occasional storms

from the east.

Because D. viridis can measure up to 145 cm in length at Île aux Goélands, we

also examined the effect of the size of D. viridis on the abundance of juvenile

A. cribrosum and urchins. We applied the same sampling procedure described above

around ten haphazardly‐selected medium‐sized (50‐70 cm) and large‐sized

(110‐130 cm) D. viridis, and observations were carried out on the western side of the

74

island on six dates in July 1998 (for each plant, data from the north and south sides of

the holdfast were pooled in the analysis).

To determine if juvenile A. cribrosum and urchins were uniformly distributed

within the area swept by D. viridis, we estimated their abundance at different

distances from D. viridis holdfasts (sampled in July and August of 1997 on the eastern

side of Île aux Goélands). We recorded the distance from the holdfast for every

A. cribrosum juvenile located within a 40‐cm radius of the holdfast. Densities were

subsequently calculated for concentric 10‐cm wide zones (numbers were adjusted for

the increasing area of the concentric zones moving away from the holdfast). Because

urchins were numerous, we measured their densities in 10‐cm sections along 15‐cm

wide belt transects running north and south from the holdfast to a distance of 60 cm

(the data from the two directions were pooled in the analysis). For these

measurements, we haphazardly selected five 50‐70‐cm‐long D. viridis plants. As our

observations showed that radius of the average area swept during wave action was

31 cm (SE = 2.8), we considered areas beyond 30 cm as outside the swept area.

Protection provided by D. viridis

We investigated the effectiveness of the Desmarestia viridis canopy in providing

protection for juvenile Agarum cribrosum from urchin grazing using two approaches.

First, in July and August 1997, we conducted an experiment on the eastern side of the

island to compare the rate at which the urchins made contact with juveniles

A. cribrosum in the absence and presence of a D. viridis canopy. For each of four trials,

we identified ten juveniles under each of four haphazardly selected 55‐65‐cm

D. viridis plants (a total of 160 juvenile A. cribrosum under 16 D. viridis) by placing

a numbered stainless steel washer next to each juvenile. We then removed two of the

four D. viridis by cutting their stipes and returned after 24, 48 and 72 h to score

whether each juvenile was in contact with an urchin.

Second, in July and August of 1998, we performed a second experiment on the

western side that was similar to the trials in 1997, except that we quantified the

75

changes in the total length of A. cribrosum juveniles in the absence or presence of

a D. viridis canopy. For each of three trials, we identified and measured the length of

ten juveniles under each of six haphazardly selected 50‐60‐cm D. viridis plants. Then,

after randomly removing three plants, we measured the length of the juveniles after

24, 48, 72 and 96 h. Subsequently, we expressed length as a percentage of initial

length for each juvenile. Mean initial length was almost the same in the treatments

with (4.9 cm, SE = 0.3) and without (4.8 cm, SE = 0.2) cover (⏐t⏐ = 0.07, df = 141,

P = 0.94).

Growth of juvenile A. cribrosum

We evaluated the effect of a Desmarestia viridis canopy on the growth of

Agarum cribrosum juveniles by comparing the changes in length of juveniles in the

presence and absence of a canopy, in areas where urchins were excluded using

fences. We first identified five A. cribrosum juveniles under the cover of each of eight

haphazardly‐selected D. viridis plants (55‐65 cm in length) on the western side of the

island. Then, after removing four of the D. viridis plants (chosen at random), we

quantified the changes in total length of the juveniles at 5‐d intervals over a 39‐d

period (3 July to 11 August 1998). Prior to 3 July, we constructed fences (2 x 2 m with

60‐cm‐high walls) around each of the eight experimental areas to exclude urchins.

The fences were constructed of 0.5‐cm mesh wire supported by PVC pipes and were

bolted to holes drilled in the bedrock. We first removed all the urchins from all fences

at the start of the experiment and then made daily visits to remove urchins that had

climbed onto the cage walls. For the analysis we calculated the length of each plant as

a percentage of initial length. Individual A. cribrosum were approximately 4‐cm‐tall at

the beginning of the experiment and tended to be slightly taller in the treatment with

(4.4 cm, SE = 0.6) than without (3.5 cm, SE = 0.5) cover (⏐t⏐ = 1.22, df = 28, P = 0.23).

The data on five A. cribrosum juveniles from each treatment were lost during the

study and were thus excluded from the analyses. Although two D. viridis plants in

the cover treatment broke away in the latter part of the experiment, they were

76

replaced (within 24 h) by attaching other D. viridis plants (with cable ties) to bolts

next to the holdfasts of the original plants.

Seasonal cycle of D. viridis

Because Desmarestia viridis is an annual alga and can only provide a certain period of

protection to Agarum cribrosum, we documented the decline in size of identified

D. viridis plants on the western side of the island in both 1998 (15 haphazardly chosen

individuals observed on 16 July, 23 August, and 25 October) and 1999 (28 plants

observed on 9 June, 2 and 18 August, and 26 October).


Split‐plot ANOVAs (Montgomery, 1991) were used to analyze the distribution of

juvenile Agarum cribrosum and urchins near Desmarestia viridis with the factors,

Side (western/eastern), Location (inside/outside), and Direction (north/south) with

one nested element (plant replicate within side). The analyses were applied to the raw

numbers of juvenile A. cribrosum whereas the data on numbers of urchins were

square‐root transformed to correct for heteroscedasticity.

To assess the impact of the size of D. viridis cover on the distribution of

juvenile A. cribrosum and urchins, we performed split‐plot ANOVAs with the factors,

Size (medium/large) and Location (inside/outside) with one nested element (plant

replicate within size). The data on numbers of juveniles were square‐root transformed

to obtain normality and homoscedasticity. Since no transformations corrected for the

heteroscedasticity detected in the raw data on numbers of urchins, the ANOVA was

also applied to the rank transformed data. Because both analyses gave similar results,

we present results from the former (raw data) following the suggestion of Conover

(1980).

We evaluated the effect of the distance from the D. viridis holdfast on the

density of juvenile A. cribrosum and urchins using one‐way ANOVAs with the factor

Distance (0‐10, 10‐20, 20‐30 and 30‐40 cm from the holdfast for juvenile A. cribrosum;

77

0‐10, 10‐20, 20‐30, 30‐40, 40‐50 and 50‐60 cm from the holdfast for urchins) applied to

the raw data for numbers of juvenile A. cribrosum, and to the square‐root transformed

data for numbers of urchins (to correct for heteroscedasticity).

To analyze the frequency of contact of urchins with juveniles, we applied

a repeated measures ANOVA (Hand & Taylor, 1987; Crowder & Hand, 1990) with

the factors, Treatment (with/without canopy) and Time (after 24, 48 and 72 h).

We treated this analysis as the particular case of the generalized linear models

(McCullagh & Nelder, 1989) that assumes a binomial distribution of the response

variable (dichotomous response, with or without contact). Thus, we did not test for

normality and homoscedasticity in the data, and no binomial variation was detected.

A second repeated‐measures ANOVA was carried out to test whether a D. viridis

cover protected juvenile A. cribrosum from urchin grazing with the factors, Treatment

(with/without canopy) and Time (after 24, 48, 72, and 96 h). Since we detected an

increase in the variance over time for each treatment and the correlation in the data

over time was different from one treatment to another, we modeled the covariance

structure with a specific heterogeneous compound symmetry structure (Proc Mixed,

Type = CSH, SAS Institute Inc., 1999). We applied the analysis to the raw data as the

data for each day were normally distributed.

A randomization (permutation) test (Legendre & Legendre, 1998) was used to

analyze the effect of a D. viridis cover on the growth of juvenile A. cribrosum during

the overall sampling period. Thus, we compared the result of a standard Student’s

t‐test obtained from our field sampling with the 95th percentile of the distribution of

Student’s t‐test generated by any possible randomized combination of our sampling

results.



distribution of the residuals and by using Levene test (Snedecor & Cochran, 1989).

Differences between levels within a factor were identified using least‐square means

multiple comparisons tests (LS means, SAS Institute Inc., 1999), with sequential

78

Bonferonni correction of probabilities (Rice, 1989; Peres‐Neto, 1999). We used

a significance threshold of 0.05 for all statistical tests.

Results

Distribution of juvenile A. cribrosum and urchins near D. viridis

The 1997 sampling inside and outside of the cover of Desmarestia viridis plants

showed inverse patterns for the densities of juvenile Agarum cribrosum and urchins

(Fig. 4.1). Juvenile A. cribrosum were six times more abundant inside than outside the

cover (P < 0.0001, Table 4.1), whereas urchins were more than two times more

abundant outside than inside the cover (P < 0.0001, Table 4.1). In this analysis, we

pooled the data from different dates on the eastern side, as a preliminary analysis

revealed no difference between the two sampling dates (F1,18 = 0.03, P = 0.87 for

juveniles; F1,18 = 0.64, P = 0.43 for urchins), and then compared them to the data from

the single sampling date on the western side.

The densities (individuals m‐2) of both A. cribrosum juveniles and urchins were

significantly greater on the western than on the eastern side of the island

(66.7 ± 73.4 [SD] vs. 41.3 ± 45.0 for juveniles, P = 0.041; 168.0 ± 70.1 vs. 81.3 ± 83.7 for

urchins, P = 0.0032). Although densities of both A. cribrosum juveniles and urchins

were similar at both sides of the island outside the cover (LS means, P = 0.80 and

P = 0.19, respectively), they were greater at the western side inside the cover

(LS means, P = 0.031 and P < 0.001, respectively). Moreover, the distribution of

urchins in the vicinity of D. viridis varied between sides, there being significant

interactions among factors (Location x Direction, Side x Location, and

Side x Direction; Table 4.1). At the western side of the island, more urchins were

found on the south than on the north side of D. viridis plants (LS means, P = 0.045),

and juveniles tended to be more abundant on the north than on the south side of the

D. viridis plants, although not significantly (80.0 ± 96.0 [SD] vs. 53.3 ± 47.1; LS means,

79

0

20

40

60

80

100

120

140

160

180

A. cribrosum S. droebachiensis

P < 0.0001

P < 0.0001Inside D. viridis

Outside D. viridis

Fig. 4.1. Mean density (+ SE) of juvenile kelp Agarum cribrosum and of the green sea urchin, Strongylocentrotus droebachiensis, inside and outside the zone swept by the filamentous phaeophyte Desmarestia viridis. P‐values are for comparisons between the two zones. For the urchin, the data have been back‐transformed from the transformed data used in the analysis.

80

Table 4.1. Summary of split‐plot ANOVAs showing the effect of the Side of the island (eastern or western sides), Location (inside or outside a canopy of Desmarestia viridis) and Direction (north or south of D. viridis) on the mean density of juvenile kelp Agarum cribrosum and of the green sea urchin, Strongylocentrotus droebachiensis. Data on urchins were square‐root transformed.

Source of variation df MS F‐value P Juvenile Agarum cribrosum Side 1 5134.22 6.24 0.041 Plant(Side) 7 822.86 0.46 0.85 Location 1 52921.00 29.37 <0.0001 Direction 1 2048.00 1.14 0.30 Location x Direction 1 6883.56 3.82 0.064 Side x Location 1 7523.56 4.17 0.054 Side x Direction 1 910.22 0.51 0.49 Side x Location x Direction 1 512.00 0.28 0.60 Error 21 1802.16 Corrected total 35 Urchins Side 1 237.19 19.37 0.0032 Plant(Side) 7 12.25 1.89 0.12 Location 1 269.43 41.63 <0.0001 Direction 1 19.11 2.95 0.10 Location x Direction 1 60.02 9.27 0.0061 Side x Location 1 83.67 12.93 0.0017 Side x Direction 1 34.16 5.28 0.032 Side x Location x Direction 1 2.74 0.42 0.52 Error 21 6.47 Corrected total 35

81

P = 0.76). Such differences were not detected on the eastern side of the island where

densities of juvenile algae and urchins were similar on north and south sides

of D. viridis (LS means, P = 0.62 and P = 0.76, respectively).

Whereas tidal currents close to the bottom generally showed a similar pattern

on both sides of the island, current speeds (summer 1999) were higher on the western

(6.79 cm s‐1) than on the eastern (4.70 cm s‐1) side (⏐t⏐ = 21.69, df = 10306, P < 0.0001),

and the strongest currents on the western side were oriented towards the

north‐northeast with a peak at 15° (Fig. 4.2).

In addition to confirming the inverse distributions of A. cribrosum juveniles and

urchins inside and outside the cover of D. viridis, the sampling around medium‐sized

and large plants revealed that there were higher densities of juveniles under large

D. viridis (LS means, P = 0.015; Fig. 4.3). The density of juveniles outside the cover did

not vary with plant size (LS means, P = 0.92; Fig. 4.3). No effect of D. viridis size was

detected for urchin densities (split‐plot ANOVA, F1,18 = 1.25, P = 0.28; Fig. 4.3).

The more detailed sampling in the vicinity of the D. viridis plants revealed an

effect of distance from the holdfast on the densities of both A. cribrosum juveniles

(ANOVA, F3,12 = 7.58, P = 0.0042) and urchins (ANOVA, F5,20 = 7.03, P = 0.0006).

The data indicated that A. cribrosum juveniles were most abundant at a distance of

10‐20 cm from the holdfast (Fig. 4.4). The juveniles were also less abundant just

outside the swept zone (30‐40 cm) than within the swept zone as a whole (Group

contrast test, F1,12 = 14.43, P < 0.01). Again, urchin densities were much greater outside

than inside the swept area (Group contrast test, F1,20 = 31.77, P < 0.0001) and within the

swept zone tended to be least abundant closest to the holdfast (0‐20 cm).

Protection provided by D. viridis

In the 1997 experiment, the observations over three consecutive days showed that the

number of Agarum cribrosum juveniles in contact with urchins increased more rapidly

82

Current direction (% frequency)

Western side Eastern side

Current speed (cm s-1)

T (°C) = 5.98 (SE=1.88)Salinity (PPT) = 31.25 (SE=0.29)

T (°C) = 6.33 (SE=1.93)Salinity (PPT) = 31.06 (SE=0.34)

510152025 90

180

270

360

90

180

270

360

510152025

90

180

270

360

2810 46 90

180

270

360

2810 46

Fig. 4.2. Frequencies distribution (%) for current direction and speed from a current meter located at 1 m above the bottom on the western and eastern sides of Île aux Goélands during the summer of 1999.

83

A. cribrosum

0

10

20

30

40

50

60

70

50-70 cm D. viridis

110-130 cm D. viridis

a

a

�b

b

Inside Outside

c c

0

50

100

150

200

250

300

b

a

b

S. droebachiensis

Fig. 4.3. Mean density (+ SE) of juvenile Agarum cribrosum and of the green sea urchin, Strongylocentrotus droebachiensis, inside and outside the zone swept by medium‐sized (50‐70 cm) and large (110‐130 cm) Desmarestia viridis. For juveniles, the data have been back‐transformed from the transformed data used in the analysis. Values not sharing the same letter are significantly different (LS means, P < 0.05).

84

0-10 10-20 20-30 30-40 40-50 50-600

20

40

60

80

100

120

140

160

180

Swept zone Non-swept zone

Distance from the D. viridis holdfast (cm)

A

A

B

AB

B

A

ab

a

b

ab

S. droebachiensis

A. cribrosum

Fig. 4.4. Mean density (+ SE) of juvenile kelp Agarum cribrosum and of the green sea urchin, Strongylocentrotus droebachiensis, at different distances from the holdfast of the filamentous phaeophyte Desmarestia viridis (A. cribrosum not sampled beyond 30‐40 cm). The limit of the area swept by D. viridis was 31 cm. For the urchins, the data have been back‐transformed from the transformed data used in the analysis. Values not sharing the same letter (small letters for juveniles, capitals for urchins) are significantly different (LS means, P < 0.05).

85

in the absence than in the presence of a Desmarestia viridis canopy (Fig. 4.5). At the

end of the experiment, over 90 % of the juveniles in the canopy‐removal treatment

had had contact with urchins compared to <15 % in the control treatment.

The 1998 experiment, which quantified changes in length of juvenile

A. cribrosum over four consecutive days in the absence and presence of a D. viridis

canopy, showed a 53 % loss in length when the canopy was removed compared to

a non significant loss (6 %) with the canopy present (Fig. 4.6). The decrease in frond

length in the absence of the cover appeared linear with time, and the difference

between treatments became significant after 3 d (LS means, P < 0.01; Fig. 4.6). At the

end of the study, 34 % of the juveniles with the removed canopy had been completely

grazed, compared to only 3 % when the canopy was left intact. The predicted time to

disappearance of A. cribrosum juveniles, based on the rates of loss recorded over 4 d

(and assuming that urchins continue to graze at the same rate), would be an order

of magnitude less in the presence (7.6 d) than in the absence (71.4 d) of D. viridis.

Growth of juvenile A. cribrosum

Although the mean growth (% increase in size) of Agarum cribrosum juveniles over the

39‐d period was greater in the absence (106.4 %, [SE = 12.6]) than in the presence

(77.4 %, [SE = 7.5]) of a Desmarestia viridis cover, the difference was not significant

(P = 0.14; n = 4).

Seasonal cycle of D. viridis

The measurements of Desmarestia viridis plants in 1998 and 1999 showed a rapid

decrease in size during the summer. Mean length decreased by 78 % between 16 July

and 23 August in 1998 (equivalent to a 2 % decrease per day) and by 91 % between

9 June and 18 August in 1999 (1.3 % per day). In both years, no D. viridis sporophytes

were observed during dives in late October. Thus, Agarum cribrosum juveniles that

developed under D. viridis became fully exposed to urchins during September and

October.

86

24 48 720

20

40

60

80

100

Time (h)

Without D. viridis

With D. viridis

a

b

c

d

d d

Fig. 4.5. Mean percentage (+ SE) of juvenile kelp Agarum cribrosum in contact with the green sea urchin, Strongylocentrotus droebachiensis, over 72 h in the presence or absence of a cover of the filamentous phaeophyte Desmarestia viridis. Values not sharing the same letter are significantly different (LS means, P < 0.05).

87

a

24 48 72 960

20

40

60

80

100With D. viridis

Without D. viridis

Time (h)

aaa

a

a

b

b

Fig. 4.6. Length as a percentage of initial length (± SE) of juvenile kelp Agarum cribrosum over 96 h, when growing under a cover of the filamentous phaeophyte Desmarestia viridis and when the cover has been removed. Values not sharing the same letter are significantly different (LS means, P < 0.05).

88

Discussion

The plant‐herbivore interaction described in this study represents an unusual marine

plant defense association in which the annual alga Desmarestia viridis provides a

short‐term shelter for juvenile stages of the kelp Agarum cribrosum from grazing by

the green sea urchin, Strongylocentrotus droebachiensis. The repulsion of sea urchins

by the movement of algal fronds (induced by water motion) has been previously

noted for D. viridis (Himmelman & Steele, 1971; Himmelman, 1984; Konar, 2000),

Desmarestia ligulata (Dayton, 1985a), and the kelp Laminaria pallida

(Velimirov & Griffiths, 1979). We showed that the cover of D. viridis had a strong

impact on the spatial distribution of both juvenile A. cribrosum and urchins, but in

opposite ways. Juvenile A. cribrosum were almost always restricted to the area swept

by D. viridis plants, and urchins were mostly located outside the swept areas.

The variations observed between our two study sites indicate that the relationship

between the repulsive effect of D. viridis on urchins and the recruitment

of A. cribrosum is sensitive to several factors. Although urchin densities outside of

areas covered by D. viridis were similar on the western and eastern sides of

Île aux Goélands, urchins were better able to invade the area swept by D. viridis on

the western side. Should grazing by urchins be responsible for the distribution of

A. cribrosum juveniles, one might predict that the densities of juveniles under

D. viridis would have been lower on the western than on the eastern side of the island

(urchins were more abundant under the alga on the western side). However, the

inverse was observed. This seemingly contradicting pattern could be due to such

factors as changes in urchin behavior related to temporal changes in water movement

(e.g., sampling during a calm period) or to higher levels of A. cribrosum recruitment

on the western side.

Variation was also seen at finer spatial scales. Urchins on the western side of

Île aux Goélands were more abundant on the south side of D. viridis, whereas

A. cribrosum juveniles tended to be most common on the north side, a pattern

consistent with the idea that recruitment success is related to grazing intensity.

89

Again, the hydrodynamic conditions might explain this variation. The strongest

subtidal currents were observed on the western side of the island and were usually

oriented in a northerly direction. The speed and orientation data suggest there may

have been prolonged periods of unidirectional flow that may have dampened the

wave‐induced oscillations in shallow water. If so, one would expect the deterrence of

urchins by D. viridis to be strongest in the northerly direction. Green sea urchins are

highly mobile animals (maximum 2.6 m h‐1, P. Gagnon, unpublished data) and

readily shift their foraging behavior in response to other organisms and to physical

conditions (Garnick, 1978; Larson et al., 1980; Johnson & Mann, 1982; Mann et al.,

1984; Scheibling & Hamm, 1991). Thus, they could have temporarily attacked the

D. viridis plants from the southern side during periods when currents push the

canopy towards the north. These results indicate the importance of considering

the behavioral repertoire of mobile consumers relative to variations in the small‐scale

physical environment.

The probability of contact between urchins and juvenile A. cribrosum appears

to be influenced by the nature of the D. viridis canopy itself. Larger D. viridis plants

appear to provide better protection from urchins as higher densities of juvenile

A. cribrosum were found beneath them. We did not detect an effect of D. viridis size on

the density of urchins, suggesting that density alone does not always predict the

potential effect of herbivory. The enhanced recruitment under large plants was

possibly because the larger plants interfered more with the urchin’s foraging.

Alternatively, larger plants with their greater swept areas may provide a buffer zone

with moderate levels of grazing around a more highly protected central zone. This

idea is supported by the micro‐distributional data that showed higher densities

of juveniles, and corresponding lower densities of urchins, in the center of the swept

area. Increased algal recruitment in areas swept by larger plants has also been

reported for the kelp Laminaria pallida in South Africa (Velimirov & Griffiths, 1979).

The increase was attributed to decreased disturbance of the bottom by the fronds,

because the stiffer stipes of the large plants held the fronds away from this zone.

Although the stipes of D. viridis do not appear to stiffen with size, the branches near

90

the holdfast appear to move less than the outer branches (possibly because wave

action is attenuated). In contrast, all branches of small plants move back and forth

with wave action. The increased number of A. cribrosum juveniles near the holdfasts

of larger D. viridis might similarly be due to a reduced level of frond abrasion.

The damage to a plant from a herbivore is a function of (1) the likelihood that

the herbivore will come into contact with the plant, (2) the probability that the contact

leads to a portion of the plant being eaten, and (3) the loss in fitness due to damage

from the herbivore (Lubchenco & Gaines, 1981). During the 1997 experiment, the

probability of contact between an urchin and an A. cribrosum juvenile was 6 times

greater in the absence than in the presence of a D. viridis canopy. Furthermore, in the

1998 experiment tissue loss was 9 times greater in the absence than in the presence of

a D. viridis canopy. Thus, a D. viridis cover reduces both the probability of contact

with urchins and also tissue loss from urchin grazing. Extension of our tissue loss

curves, assuming the levels of interactions are similar over a longer period, suggests

that A. cribrosum juveniles would survive an order of magnitude longer in the

presence than in the absence of a D. viridis canopy. Van Alstyne (1999) showed that

many herbivores prefer the juvenile stages of algae, perhaps because chemical

or morphological defenses are less well developed. Possibly juvenile A. cribrosum

undergo chemical or structural changes to increase resistance to grazing as they age.

Adult A. cribrosum are not preferred foods of urchins (Vadas, 1977; Larson et al., 1980;

Himmelman & Nédélec, 1990). Thus, an association with D. viridis may provide

A. cribrosum juveniles with an escape from grazing during the critical period when

their defenses are not yet well developed.

In addition to affecting urchin behavior, macrophyte canopies can also reduce

(1) flow (Ackerman & Okubo, 1993; Hurd et al., 1997), and thus particle transport

(Eckman et al., 1989) and nutrient exchange (Gerard, 1982), and (2) light availability

(Gerard, 1984; Reed & Foster, 1984). Reduced light and nutrients could reduce growth

rates of plants growing under an algal canopy. For example, a cover of the brown

alga Sargassum filipendula has been shown to decrease the growth of the associated

91

red alga Gracilaria tikvahiae (Pfister & Hay, 1988). Vadas (1968) has shown that growth

of A. cribrosum is sensitive to reduced light intensities, which is consistent with our

39‐d experiment in which the growth rate of juvenile A. cribrosum was 29 % less

under a D. viridis canopy than with the canopy removed. This decrease was not

significant but might have been if the experiment had run for a longer period or

sample size was increased. Nevertheless, on urchin barrens, the costs of loss in

growth associated with living under a canopy are clearly outweighed by the benefits

of reduced urchin grazing.

Most studies of associational plant defenses have described the results of

long‐term interactions between species. In contrast, the association of A. cribrosum

with D. viridis is short‐lived because of the annual life cycle of D. viridis.

Our observations in 1998 and 1999 show that the length of D. viridis plants decreases

by about 1.5 % daily during the summer, and that all the plants are gone by late

October. Thus, the protection provided by D. viridis diminishes over the summer and

disappears in the autumn. Nevertheless, two factors may favor the survival of the

remaining juveniles in the autumn. The first is the decrease in feeding rates

of the urchins during late summer and autumn (Vadas, 1977; Larson et al., 1980;

Meidel & Scheibling, 1998). The second is that juveniles may switch from an escape to

a defensive strategy once they have reached a size where their own defenses are

developed (e.g., more polyphenolics, tougher tissues, and a size where the sweeping

action of its own blade might deter urchin attacks).

The interaction that we examined appears to represent a successional step

leading to the formation of mature stands of A. cribrosum. Thus, D. viridis colonizes

and grows rapidly during the winter, and then provides areas of decreased urchin

grazing where vulnerable A. cribrosum juveniles can become established. Then, the

A. cribrosum juveniles may be able to survive on their own when D. viridis disappears

in the autumn. Such an interaction would represent successional facilitation

(sensu Connell & Slatyer, 1977), which has been receiving increasing recognition as a

key process structuring communities (Walker et al., 1986; Wood & del Moral, 1987;

92

Callaway, 1995; Bruno & Bertness, 2001). Long‐term studies are required to determine

the survival rate of young A. cribrosum once the D. viridis cover disappears.

Such studies should also evaluate the influence of such factors as topographical

heterogeneity, ice scouring (which may destroy or disperse urchins in localized areas)

and wave action.

Chapitre V

Multiple plant defenses against herbivores: the interplay of chemical, morphological, and hydrodynamic factors

94

Résumé

Les plantes disposent d’un large éventail de mécanismes pour repousser les

herbivores. Même si les défenses chimiques sont bien documentées dans

les communautés terrestres et marines, peu d’études ont examiné les interactions

possibles entre la répulsion chimique et les autres types de défense. Nous avons

réalisé des expériences en laboratoire afin de déterminer si la résistance de

Desmarestia viridis (algue brune fortement acide et délicatement branchue) au

broutage de l’oursin s’explique par ses propriétés chimiques ou par une défense

mécanique attribuable à sa morphologie. Lors d’expériences de chémodétection en

utilisant un labyrinthe en Y, les oursins étaient attirés par des frondes de D. viridis

dans certains essais, mais repoussés dans d’autres (la température variait entre les

différents essais). Au contraire, les oursins étaient ni attirés, ni repoussés par leur

nourriture préférée, la laminaire Alaria esculenta. Des tests dans un générateur de

vagues ont démontré que l’agrégation des oursins près ou sur D. viridis diminuait

avec l’augmentation du mouvement de balayage provoqué par l’augmentation du

mouvement de l’eau. Cette habileté de D. viridis à repousser les attaques de l’oursin

décroissait drastiquement lorsque la densité de l’oursin atteignait un niveau critique.

Ces résultats sont comparables à nos observations de terrain à l’effet que les oursins

attaquent facilement D. viridis lorsque les conditions sont calmes mais évitent le

contact avec l’algue lorsqu’agitée par l’action des vagues, mêmes modérées.

Contrairement aux spéculations littéraires que l’acidité des Desmarestiales procure

une protection contre le broutage, nos résultats démontrent que la composition

chimique de D. viridis est à elle seule ni nécessaire, ni suffisante pour protéger l’algue

contre les oursins. Nous démontrons que la protection de D. viridis contre les

herbivores dépend principalement de l’interaction entre sa morphologie branchue et

l’action des vagues.

95

Abstract

Plants rely on a wide array of mechanisms to deter herbivores. Although chemical

defenses are well documented in terrestrial and marine communities, few studies

have examined possible interactions between chemical deterrence and other types of

defense. We conducted laboratory experiments to determine if the resistance of the

highly acidic, delicately‐branched, brown alga Desmarestia viridis to urchin grazing is

explained by its chemical properties or by mechanical defenses related to its

morphology. In Y‐maze chemodetection experiments, urchins were attracted by

D. viridis fronds in some trials but repelled in others (temperature varied among

trials). In contrast, urchins were neither attracted to nor repelled by their preferred

food, the kelp Alaria esculenta. Trials in a wave tank showed that the aggregation of

urchins near or on D. viridis decreased as the sweeping motion of D. viridis increased

with increased water motion. This ability of D. viridis to repel urchin attacks

decreased markedly when urchin density attained a critical level. These results are

consistent with field observations that urchins readily attack D. viridis under calm

conditions but avoid contact with the alga when it is moved about even by moderate

wave action. Contrary to speculation in the literature that the acidic nature of

Desmarestiales provides protection from grazers, our results show that the chemical

make‐up of D. viridis alone is neither necessary nor sufficient to protect the alga from

urchins. We show that the protection of D. viridis from herbivores depends primarily

on the interaction between its highly‐branched morphology and wave action.

96

Introduction

Herbivory is ubiquitous in terrestrial and marine communities, and the substantial

increase in studies on plant‐herbivore interactions over the last three decades

(see reviews in Lubchenco & Gaines, 1981; Hay & Fenical, 1988; Lindroth, 1989;

Coley & Barone, 1996; Zamora et al., 1999) has lead to a better understanding of

processes regulating populations and communities (Dayton, 1985b; Karban & Myers,

1989; Sousa & Connell, 1992; Callaway et al., 2000; Schmitz et al., 2000). In response to

herbivory, plants have evolved a variety of defenses including spatial or temporal

escapes (Atsatt & OʹDowd, 1976; Hay, 1986), morphological defenses (Duffy & Hay,

1990; Hay et al., 1994; Lucas et al., 2000), and chemical deterrents (Hay & Fenical,

1988; Paul & Van Alstyne, 1992). Most recent studies have focused on chemical

defenses (e.g., Pavia & Toth, 2000; Taylor et al., 2002), perhaps to the neglect of other

mechanisms and the possibility of multiple plant defenses (Hay et al., 1994).

Herbivory can be intense in marine benthic communities (Lubchenco &

Gaines, 1981), especially in tropical reefs (Carpenter, 1986; Lewis, 1986) and nearshore

temperate habitats (Breen & Mann, 1976b; Scheibling et al., 1999; Gagnon et al., in

press) where intensive grazing by fishes and sea urchins can lead to the formation of

“barren zones” (Lawrence, 1975). Obvious mechanical defenses, such as calcified

tissues, have long been recognized as a key adaptation to living in heavily grazed

environments (Steneck, 1986), but more recent studies have demonstrated that

chemical defenses are also common in marine plants and can be highly effective in

deterring herbivores (Duffy & Hay, 1990; Bolser & Hay, 1996; Schnitzler et al., 2001).

Brown algae of the order Desmarestiales, common in temperate coastal waters

(Peters et al., 1997), are unusual because of their production and storage of high

concentrations of sulfuric acid (H2SO4) within cell vacuoles, resulting in extremely

low internal pH (down to ~0.5 pH, McClintock et al., 1982). A number of studies have

suggested this trait chemically repulses grazers (Anderson & Velimirov, 1982;

Dayton, 1985a; Himmelman & Nédélec, 1990). This idea, although attractive, has only

recently received experimental support: the addition of sulfuric acid to an agar‐based

97

food source reduced urchin feeding rates in the laboratory (Pelletreau

& Muller‐Parker, 2002). However, field observations show that urchins feed readily

on Desmarestia sp. under certain environmental conditions (Himmelman & Steele,

1971; Cowen et al., 1982; Konar, 2000). Thus, laboratory studies might not reflect how

grazers respond to acidic species of Desmarestia, as they do not incorporate important

environmental factors. Water motion and temperature, urchin density and motivation

to feed likely affect the behavior of urchins (Garnick, 1978; Bernstein et al., 1983;

Mann et al., 1984; Vadas et al., 1986; Kawamata, 1998) and could modify the efficacy

of chemical defenses.

In the Mingan Islands, northern Gulf of St. Lawrence, the green sea urchin,

Strongylocentrotus droebachiensis, dominates bedrock platforms, and, as in many other

regions, its grazing leads to the formation of urchin barrens (Lawrence, 1975). This

community state is strikingly different from adjacent shallow zones where urchins

are scarce and kelp are abundant (primarily Alaria esculenta). The causal nature of this

pattern is well established both in the Mingan Islands (Gagnon et al., in press) and

elsewhere (e.g., Pearse & Hines, 1979; Himmelman et al., 1983; Witman, 1987;

Scheibling et al., 1999): removal of urchins leads to rapid colonization by kelp and

other algae. Desmarestia viridis is an exception to this pattern, being one of the few

fleshy macrophytes that can establish on urchin barrens. It is an annual alga that

occurs exclusively on barrens, either individually or in small groups (Fig. 5.1). It has

a small holdfast, a thin highly‐flexible stipe, and a profusely‐branched, filamentous

frond that can measure up to 1.5 m in length.

The occurrence of this fleshy macrophyte in an intensively‐grazed

environment, along with its known high acidity (McClintock et al., 1982), suggests

that it deters grazing using chemical defenses. Several observations support this idea.

First, D. viridis is only consumed at low rates when offered to urchins in the

laboratory (Himmelman & Nédélec, 1990). Second, field experiments comparing the

degree to which urchins move towards monospecific bundles of various algae reveal

98

Fig. 5.1. An isolated plant of Desmarestia viridis (~60 cm in length) surrounded by a dense aggregation of the green sea urchin, Strongylocentrotus droebachiensis, at 5 m in depth on a gently‐sloped bedrock platform in the Mingan Islands in June 2000.

99

that urchins are not attracted to D. viridis (Himmelman & Nédélec, 1990) when other

algal species are present. Third, the consumption rate of D. viridis is generally much

lower than for preferred algae such as the kelps A. esculenta and Laminaria spp.

(Vadas, 1977; Larson et al., 1980; Himmelman, 1984; Himmelman & Nédélec, 1990).

Although the evidence for chemical deterrence is convincing, other

possibilities have not been adequately explored. The delicate structure of D. viridis

permits the frond to sweep back and forth over the bottom with even slight wave

action, and observations and experiments indicate that urchins are much less

abundant under a D. viridis canopy (Himmelman, 1984; Konar, 2000; Gagnon et al.,

2003b) as well as for other algae (Velimirov & Griffiths, 1979; Dayton, 1985a).

Although the low number of urchins under Desmarestia spp. has been interpreted as a

response to the chemicals in the fronds (e.g., ʺacid broomingʺ sensu Dayton, 1985a),

the repulsive effect may be due simply to the mechanical contact between algal

fronds and urchins (Konar, 2000). This is likely the case for the kelp Laminaria pallida

(Velimirov & Griffiths, 1979) that has no known chemical defenses.

Here, we test whether the ability of D. viridis to resist urchin grazing depends

on the interaction between its morphology, hydrodynamics and urchin abundance, or

relies on its chemical make‐up. Our objectives are to (1) compare the degree to which

urchins are attracted towards D. viridis and the preferred alga A. esculenta,

(2) quantify the accumulation of urchins on and around D. viridis under different

levels of wave action, and (3) determine whether urchin density affects the degree to

which urchins accumulate on D. viridis.


Our study was conducted during the summers of 1999 and 2001 at Havre‐Saint‐Pierre

in the Mingan Islands in the northern Gulf of St. Lawrence, eastern Canada

(50°13´6˝N, 63°41´12˝W). Experiments were performed in the laboratory with the

green sea urchin, Strongylocentrotus droebachiensis (2 to 6 cm in diameter), and the

100

algae Desmarestia viridis and Alaria esculenta, which were all collected at depths of

3‐8 m around Île aux Goélands and Île du Havre (Fig. 2.2), two islands located

approximately 5 km from the laboratory, and then maintained in tanks continually

supplied with seawater pumped from 10 m in depth.

Response of urchins to odors of Desmarestia viridis

To examine whether Desmarestia viridis chemically repulses Strongylocentrotus

droebachiensis, we compared urchin displacement towards chemical odors released by

D. viridis and the preferred alga, Alaria esculenta, in a Y‐maze (Castilla & Crisp, 1970).

The experiment included eight treatments of which six were conducted between

3‐6 °C: (1) a control for asymmetries in the maze (flowing seawater only); (2) a control

for the presence of conspecifics (6 urchins measuring 4 cm in test diameter); and

treatments with algal stimuli, (3) 75‐g piece of an A. esculenta frond; (4) 250‐g piece of

an A. esculenta frond; (5) 75‐g piece of a D. viridis frond; (6) 250‐g piece of a D. viridis

frond. The two other treatments, (7) a control for asymmetries in the maze (flowing

seawater only) and (8) 250‐g piece of a D. viridis frond, were conducted at higher

temperatures (10‐11 °C) as temperature can affect the displacement and

chemodetection ability of the green sea urchin (Mann et al., 1984). Our system did not

allow us to control water temperature, but we could take advantage of natural

variations.

Each trial began with the addition of the stimulus to one arm (45 cm long and

17 cm wide) of the Y‐maze followed 30 s later by the addition of 10 urchins

(4‐cm diameter) to the basal portion of the Y‐maze (20 cm long and 17 cm wide). This

sequence allowed odors to spread towards the mixing zone by the flow of seawater

(2.4 L min‐1) running through perforations positioned at 0.5 cm above the bottom in

the end of each arm. The water from the two arms mixed just below the point where

the arms joined and left the maze through perforations at 5.0 cm above the bottom at

the end of the basal portion, therefore maintaining a 5‐cm depth throughout the

maze.

101

The displacement of the urchins into the arms of the maze was evaluated every

10 min during a 40‐min period. Once observed in an arm, urchins were removed to

avoid interference with other urchins. At the end of each trial, those urchins that had

remained in the basal arm were considered to have not made a choice. Time and

logistic considerations limited us to 4 trials of the same treatment per day. Each

treatment was repeated 12 times (= 3 blocks x 4 trials) between 23 June and 17 August

of 1999. To eliminate possible biases due to asymmetries in the maze or in the

laboratory (e.g., light), we added the stimulus to a randomly‐selected arm of the maze

at the start of each day and then alternated the arm into which the stimulus was

added for subsequent trials. Between trials, the maze was drained and wiped with

a cotton towel.

Mechanical repulsion of urchins by Desmarestia viridis

To examine the degree to which the wave‐induced movement of Desmarestia viridis

fronds mechanically repulses sea urchins, we quantified the success of attack by

urchins on D. viridis or an algal mimic (defined here as the number of urchins in

contact with the alga or the mimic) in five treatments in a wave tank: (1) in the

absence of D. viridis and waves (control), (2) in the presence of a cotton mop without

waves to measure the tendency of urchins to aggregate on a chemically‐neutral,

D. viridis‐like mimic, (3) in the presence of D. viridis without waves to quantify the

vulnerability of D. viridis in the absence of wave action, (4) in the presence of D. viridis

with waves that caused a 5‐cm sweep of the fronds, and (5) in the presence of

D. viridis with waves which produced a 10‐cm sweep of the fronds. Sweep was

defined as the marginal extension of the area touched by the plant, and not the total

displacement of any particular part of the alga.

The wave tank has been described previously (see Gagnon et al., 2003a; but

also Appendix 5.1). A pulley system allowed us to control the back and forth water

motion and thus the amplitude of the sweeping action of D. viridis. The tank bottom

measured 1.94 x 1.22 m and was completely covered with concrete tiles with their

surface sculpted with undulations, holes and cracks to simulate the irregularities of

102

the natural substratum. Each corner of the tank was illuminated by a 150 W floodlight

positioned 90 cm above the bottom. For treatments with D. viridis or the algal mimic,

either a new plant (50‐60 cm in length) or a mop head with 50‐cm long cotton

filaments was attached with a plastic cable tie to an eye‐bolt at the center of the tank.

For the treatments with water motion, the system was set up to generate

23 wave cycles min‐1, equivalent to the wave frequency under moderate winds

(P. Gagnon, personal observation). To manipulate the displacement of the algal frond,

the height of the water column was adjusted to alter the strength of the waves

(deeper water produced stronger waves).

Urchins (2 to 6 cm in test diameter) were collected in July 2001 near

Île aux Goélands, and then haphazardly split into 5 groups of 275 individuals that

were fed every third day with the kelp Alaria esculenta until the end of the

experiments. The trials were run over 5 consecutive days, starting 2 d after

collection (sea water temperature: 1.5 to 5.5 °C), and the 5 groups of urchins were

subjected each day to one of the 5 treatments in a Latin square design.

Desmarestia viridis plants were freshly collected from near Île aux Goélands the

morning of each trial and were kept continually immerged in plastic buckets

provided with a constant inflow of seawater to avoid the release of chemical

compounds induced by stagnation of the water or by exposition of the fronds to air.

Mop heads were soaked for 24 h to minimize possible release of odors from the

cotton filaments. To explore possible site differences, a second group of urchins was

collected two weeks later near Île du Havre and subjected to an identical second

series of trials. Unexpectedly, water temperatures changed significantly between the

two experiments, and thus the effect of site could not be separated from an effect of

temperature.

In each of 50 trials, the urchins were first evenly distributed over the entire

tank bottom at a density of 116 m‐2, equivalent to an intermediate density of urchins

on barrens in the Mingan Islands (Gagnon et al., 2003b; Gagnon et al., in press). After

20, 40 and 60 min, we counted the number of urchins in contact with D. viridis or the

103

algal mimic and the number of urchins on the walls of the tank. We interpreted the

number of urchins on the walls as an index of the tendency of urchins to move within

the tank. For those treatments with water motion, the urchins were allowed 1 min to

attach themselves to the bottom before the motor was turned on. Prior to each new

trial, the tank was drained, cleaned and refilled with fresh seawater.

Effect of urchin density on attack success on Desmarestia viridis

We explored the importance of urchin density in determining the frequency of urchin

attacks on Desmarestia viridis by quantifying attacks under three densities, 78 m‐2,

136 m‐2 and 194 m‐2. These trials were run with water motion that produced a 5‐cm

sweep of the D. viridis fronds. On each of six sampling days (in late July and in early

August), urchins (2 to 6 cm in diameter) were collected near Île aux Goélands and

split into 3 groups to provide the necessary densities. The urchins were maintained in

running seawater for 3‐5 h prior to the trials. The procedure was as described above

(in the treatment with D. viridis with a 5‐cm sweep), except that each group of urchins

was used for just one trial (order chosen at random). Water temperatures varied

between 3 and 7 °C.

Field observations

To compare observations in our wave tank with those in the field, we counted the

numbers of urchins in contact with Desmarestia viridis plants at times of contrasting

hydrodynamic conditions at Île aux Goélands. The urchins in contact with

8 haphazardly‐chosen D. viridis plants (37‐61 cm in length) were sampled when water

motion was moderate (slack tide and ~29 km h‐1 winds), and again 4 d later when

there was low water motion (slack tide and ~9 km h‐1 winds). These conditions caused

the plants to sweep a distance of 10‐15 and <2 cm, respectively.


We used a three‐way ANOVA to analyze the tendency of the urchins to move

towards various chemical stimuli with the factors of Block (3), Treatment (8), and

104

Arm (3). We did not test for normality and homoscedasticity in the data as we treated

the analysis as a particular case of the generalized linear models (McCullagh

& Nelder, 1989; Proc Genmod, SAS Institute Inc., 1999) that assumed a binomial

distribution of the response variable (presence or absence of urchins in each arm).

Repeated measures ANOVAs (Hand & Taylor, 1987) with two nested elements

(Day of sampling within Site where urchins were collected and Group of urchins

within Site where urchins were collected) were used to examine the effect of water

motion and algal movement on (1) the attack success of urchins on Desmarestia viridis

and (2) the tendency of urchins to move within the wave tank. The factors were Site

(urchins collected from Île aux Goélands or Île du Havre), Day (10 sampling days),

Group (10 groups of urchins), Treatment (4 for the analysis on the attack success and

5 for the tendency of urchins to move) and Time (after 20, 40 and 60 min). Since

variances in the first analysis were similar over time within any of the treatments and

the correlation in the data over time was similar for all treatments, we used a pooled

covariance structure in the ANOVA (Proc Mixed, Type = CS, SAS Institute Inc., 1999).

However, because we detected fluctuations in the variances over time in the second

analysis, we modeled the covariance structure with a heterogeneous compound

symmetry structure (Proc Mixed, Type = CSH, SAS Institute Inc., 1999). We applied

the first analysis to the square‐root transformed data to obtain normality in the

distribution of the data for each observation time, whereas the second analysis was

applied to the raw data, which were normally distributed for each observation time.

The effects of urchin density on (1) the attack success of urchins on D. viridis

and (2) the tendency of urchins to move within the wave tank were analyzed using

repeated measures ANOVAs with the factors Block (6), Density (low, intermediate

and high urchin density), and Time (after 20, 40 and 60 min). Since we detected

fluctuations in the variances over time in both analyses, we modeled the covariance

structure with a heterogeneous compound symmetry structure (Proc Mixed,

Type = CSH, SAS Institute Inc., 1999). We applied both analyses to the raw data as the

data for each observation time were normally distributed.

105



distribution of the residuals and by using Levene test (Snedecor & Cochran, 1989).


multiple comparisons tests (LS means, SAS Institute Inc., 1999). A significance

threshold of 0.05 was used for all statistical tests.

Results

Response of urchins to odors of Desmarestia viridis

Our Y‐maze experiment examining the chemotaxic responses showed that under

certain circumstances urchins discriminated odors (Tables 5.1 and 5.2). The similar

numbers of urchins in the two arms in the control treatment (without stimuli at two

temperatures) indicated that there were no substantial asymmetries in the system

(Table 5.2). However, the smaller proportion of the urchins that remained in the basal

portion of the maze at the higher temperature (19.2 % at 10.3 °C versus 41.7 % at

4.7 °C; LS means, P = 0.0002) suggested that urchin displacements increased with

increasing temperature. There was no evidence of an effect of conspecifics (Table 5.2),

but the lower proportion of urchins that remained in the basal portion compared to

the test without conspecifics at 4.7 °C (LS means, P = 0.0054) suggested that odors

from conspecifics may also increase the tendency of urchins to move.

The response to chemical odors from algae provided variable results

depending on the algal species. Urchins had no reaction to Alaria esculena, but they

were either repelled or attracted to Desmarestia viridis, depending on the exact

treatment. Nearly 1.5 times as many urchins moved into the arm containing 75 g

of D. viridis as into the empty arm (LS means, P = 0.022; Table 5.2), demonstrating an

attraction. In contrast, the response to larger pieces of D. viridis (250 g) was the

opposite, and there was even greater repulsion at a higher temperature.

106

Table 5.1. Summary of the three‐way ANOVA (generalized linear model) examining the effect of Block (3 blocks of 4 replicates each), Treatment (8 stimuli), and Arm (3) on the chemotaxic response of urchins Strongylocentrotus droebachiensis to various odors (the algae Desmarestia viridis and Alaria esculenta and other conspecifics) in the Y‐maze experiment.

Source of variation df χ2 P

Block 2 1.22 0.57

Treatment 7 8.22 0.31

Block x Treatment 14 5.77 0.97

Arm 2 120.46 <0.0001

Block x Arm 4 3.61 0.46

Treatment x Arm 14 150.74 <0.0001

Block x Treatment x Arm 28 51.68 0.0042

107

Table 5.2. Directional response of the urchin Strongylocentrotus droebachiensis to various chemical stimuli in a Y‐maze choice chamber. P‐values are from statistical comparisons between the proportions of urchins that moved to the stimulus arm and the empty arm or to the two empty arms in the control tests without stimulus (LS means, P < 0.05).

Treatment Proportion of urchins (%)

Stimulus

Quantity

(g)

Sea water

temperature

(SD) (°C)

Basal

portion

Stimulus

arm

Empty

arm

P

None ‐ 4.7 (1.3) 41.7 26.7 31.6 0.41

None ‐ 10.3 (1.0) 19.2 40.0 40.8 0.89

6 urchins ‐ 5.3 (0.5) 5.0 45.8 49.2 0.59

A. esculenta 75 5.5 (0.7) 17.5 45.8 36.7 0.16

A. esculenta 250 5.0 (0.0) 20.0 42.5 37.5 0.41

D. viridis 75 5.8 (0.7) 30.0 41.7 28.3 0.022

D. viridis 250 3.8 (1.7) 28.3 25.0 46.7 0.0005

D. viridis 250 11.1 (0.6) 10.8 25.8 63.4 <0.0001

108

Mechanical repulsion of urchins by Desmarestia viridis

Our wave tank trials showed that the number of urchins in contact with Desmarestia

viridis decreased sharply with an increase in the amplitude of the movement of the

alga (Table 5.3; Fig. 5.2). In trials with urchins from Île aux Goélands, the overall

number of urchins (the mean for the three 20‐min intervals) attacking a motionless

D. viridis was more than an order of magnitude greater than for D. viridis with a 5‐cm

sweep, and two orders of magnitude greater than for D. viridis with a 10‐cm sweep.

Under still conditions, D. viridis was attacked much more than the algal mimic

(LS means, P = 0.0029), suggesting that urchins were attracted to D. viridis. Our

measurements at 20‐min intervals indicated that the cumulative number of urchins

on D. viridis increased only slightly after the first 20 min and then showed no further

change with an increase in wave action and the sweeping motion of the alga (Fig. 5.2).

Although the same pattern applied to the trials with urchins from Île du Havre, the

attacks on both mimic and natural D. viridis were much greater (Fig. 5.2). The overall

number of urchins in contact with both non‐sweeping D. viridis and the algal mimic

was 2 to 3‐fold greater for urchins from Île du Havre than for urchins from Île aux

Goélands (LS means, P < 0.0001). Whereas urchins from Île aux Goélands were

virtually unable to attack sweeping fronds of D. viridis, those from Île du Havre had

some success in attacking D. viridis with a 5‐cm sweep. The lower water temperatures

during the trials with urchins from Île aux Goélands (3.6 ± 1.4 °C) relative to those

with urchins from Île du Havre (8.0 ± 0.9 °C) could have contributed to the

differences seen between the two groups of urchins.

The numbers of urchins on the walls of the tank indicated that urchin

displacement varied markedly among treatments (Table 5.3; Fig. 5.2). The numbers of

urchins on the walls were similar in all three treatments without waves, but

decreased markedly with increasing water motion (Fig. 5.2). Urchins from both Île

aux Goélands and Île du Havre were >9 times more numerous on the walls in the

trials with non‐sweeping D. viridis than in the trials with 10‐cm‐sweeping D. viridis.

109

Table 5.3. Summary of repeated measures ANOVAs (mixed linear model) showing the effect of the Site where the urchins Strongylocentrotus droebachiensis were collected (Île aux Goélands and Île du Havre), Day of sampling (10 days), the Group of urchins being tested (10 groups), Treatment (4 in the first analysis and 5 in the second), and Time (observations at 20, 40 and 60 min) on the number of urchins in contact with Desmarestia viridis (or algal mimic) and on the walls of the wave tank. Data from the first analysis were square‐root transformed whereas the raw data were used in the second analysis.


df (numerator)

df (denominator)

F‐value

P

In contact with D. viridis Site 1 16 77.66 <0.0001 Day(Site) 8 16 0.97 0.49 Group(Site) 8 16 0.90 0.54 Treatment 3 16 74.42 <0.0001 Time 2 64 49.48 <0.0001 Site x Treatment 3 16 7.80 0.0020 Site x Time 2 64 2.13 0.13 Treatment x Time 6 64 3.56 0.0042 Site x Treatment x Time 6 64 1.26 0.29

On the walls of the tank Site 1 24 0.26 0.62 Day(Site) 8 24 1.84 0.12 Group(Site) 8 24 0.90 0.53 Treatment 4 24 22.89 <0.0001 Time 2 80 132.50 <0.0001 Site x Treatment 4 24 0.18 0.94 Site x Time 2 80 0.78 0.46 Treatment x Time 8 80 15.65 <0.0001 Site x Treatment x Time 8 80 1.31 0.25

110

0

10

20

30

40

50

20 min

40 min

60 min

0

20

40

60

80

100

In contact with D. viridis

On the walls of the tank

No alga

or waves

Mimic

no

sweep

D. viridis

no

sweep

D. viridis

5-cm

sweep

D. viridis

10-cm

sweep

c

b

d d3.5 (1.0)

10.2 (1.9)

0.6 (0.2) 0.1 (0.1)

8.4 (1.2)

32.9 (3.7)

9.6 (1.3)

0.8 (0.3)

a

bb

d

48.9 (5.9) 53.4 (5.1)40.8 (5.5)

13.8 (1.9)

4.5 (0.5)

A AA

B

B

57.2 (7.9)

55.3 (7.3)

38.6 (4.2)

16.9 (2.9)

4.1 (0.6)

A

A

A

B

B

No alga

or waves

Mimic

no

sweep

D. viridis

no

sweep

D. viridis

5-cm

sweep

D. viridis

10-cm

sweep

Île aux Goélands Île du Havre

Fig. 5.2. Mean number (+ SE) of urchins Strongylocentrotus droebachiensis in contact with Desmarestia viridis or an algal mimic, and on the walls of the tank, during three consecutive 20‐min intervals in trials in a wave tank with urchins from two sites examining the effect of the amplitude of the sweeping motion of D. viridis. Mean values for the three 20‐min intervals (± SE) are shown above bars. Treatments not sharing the same letters are different (LS means tests, P < 0.05).

111

The measurements at 20‐min intervals indicated that numbers tended to level off

rapidly with increasing water motion (Fig. 5.2), and urchins were attached most

firmly at the end of trials in the 10‐cm sweep treatment (P. Gagnon, personal

observation). Finally, the absence of significant effects of either Day(Site)

or Group(Site) in the analysis on the effect of water motion and algal movement

indicated the repeated use of the same 5 groups of urchins did not affect the ability

of urchins to attack D. viridis or climb onto the walls (Table 5.3).

Effect of urchin density on attack success on Desmarestia viridis

Our wave tank trials at a constant 5‐cm sweep of Desmarestia viridis showed that the

number of urchins attacking plants increased with urchin density, but not in a linear

fashion (Table 5.4; Fig. 5.3). The numbers (means for the three 20‐min intervals) were

roughly the same at low and intermediate densities (LS means, P = 0.83), but were

more than 2‐ to 3‐fold greater at the highest density. The number of urchins attacking

D. viridis increased more rapidly during the high‐density trials than at either low or

intermediate densities. Indeed, the number of urchins attacking after 20 min in the

high‐density treatment was nearly identical to that attained after 60 min at low and

intermediate densities (Fig. 5.3).

Field observations

Our field observations of urchins in contact with Desmarestia viridis plants indicated

that the urchins were much more abundant on plants with a small (<2 cm) than

a larger (10‐15 cm) sweep, with 17.8 (SD = 2.9) and 0 urchins per plant, respectively,

a pattern highly consistent with our observations in the wave tank experiments. Most

of the urchins in contact with D. viridis had filaments of D. viridis in their jaws.

Discussion

Most studies of marine algal defenses against herbivory have described the effect of

112

Table 5.4. Summary of repeated measures ANOVAs (mixed linear model applied to raw data) showing the effect of Block (6), Density of urchins Strongylocentrotus droebachiensis (low, intermediate and high), and Time (observations at 20, 40 and 60 min) on the number of urchins in contact with Desmarestia viridis.


df (numerator)

df (denominator)

F‐value

P

Block 5 10 0.86 0.54 Density 2 10 9.12 0.0056 Time 2 30 34.54 <0.0001 Density x Time 4 30 6.15 0.0010

113

0

10

20

30

40

Low Intermediate High

a

b

ccc

c

ccd d

Urchin density

7.2 (1.2) 6.4 (1.2)

19.1 (2.5)

20 min

40 min

60 min

Fig. 5.3. Mean number (+ SE) of urchins Strongylocentrotus droebachiensis in contact with Desmarestia viridis during three consecutive 20‐min intervals in trials in a wave tank at low (78 m‐2), intermediate (136 m‐2) and high (194 m‐2) urchin densities. Mean values for the three 20‐min intervals (± SE) are shown above bars. Bars not sharing the same letters, and treatments not bracketed by the same horizontal line (at the bottom) are different (LS means tests, P < 0.05).

114

single defensive traits, therefore neglecting possible additive or synergistic effects of

multiple defenses (Hay, 1996). Our study of the highly acidic brown alga Desmarestia

viridis to grazing by the green sea urchin, Strongylocentrotus droebachiensis, showed

that the chemical make‐up of the alga provides only limited protection from grazing.

Rather, the survival of D. viridis in intensively‐grazed subtidal environments depends

on the interaction between its delicate, finely‐branched structure and the physical

environment. Waves cause fronds to sweep back and forth over the bottom and

maintain urchins at a distance. This finding contradicts the perception that chemical

defenses are paramount in protecting acidic Desmarestiales species.

Sea urchins are highly selective feeders that have been reported to detect food

over a distance of several meters (Vadas, 1977; Garnick, 1978; Larson et al., 1980;

Mann et al., 1984). Multiple‐algae choice experiments have suggested that

Desmarestia spp. are avoided by urchins (Himmelman, 1984; Pelletreau & Muller‐

Parker, 2002). However, feeding experiments and field observations show that

urchins will consume Desmarestia species in the absence of alternative algal foods

(Himmelman & Steele, 1971; Cowen et al., 1982; Himmelman & Nédélec, 1990; Konar,

2000). Our Y‐maze study demonstrated that D. viridis can even attract urchins,

although this was not the case when there was a large quantity of the alga.

Unexpectedly, we found no chemotaxic response of urchins to odors of the preferred

food, Alaria esculenta.

Temperature is likely a key factor affecting the foraging behavior of the green

sea urchin as it appears to affect both its ability to respond to odors and the rate of

displacement. In the Mingan Islands, seawater temperatures usually oscillate around

6 °C during the summer (Gagnon et al., 2003b), although they rise to above 10 °C for

several days on two or three occasions each year (J. Himmelman and coworker’s

personal observations). Urchin behavior varied with temperature in the control trials

in our Y‐maze study, and we observed increased displacements at higher

temperatures as reported for green sea urchins (Mann et al., 1984). The two

treatments in which urchins avoided odors of D. viridis were those conducted near

115

the two summer temperature extremes for the Mingan Islands (3.8 and 11.1 °C).

At 5.0‐5.8 °C, urchins were attracted to D. viridis, but not to their preferred food,

Alaria esculenta. Repulsion of urchins by D. viridis at higher temperatures could be

related to increased acid release, but this is not likely the case at low temperatures.

Avoidance of algae by green sea urchins at low temperatures was reported by

Mann et al. (1984). Attractiveness towards a preferred alga (Laminaria longicruris)

declined with temperature and urchins were even repulsed at 2 °C. These

inexplicable results might be related to changes in the behavior and physiology of the

urchin at low temperatures.

Although there is indication that water motion can affect the behavioral

repertoire of animals (e.g., displacement and feeding rates, Rochette et al., 1994;

Kawamata, 1998; Gagnon et al., 2003a; Siddon & Witman, 2003), including behaviors

related to predation, herbivory, and competition (Cowen et al., 1982; Witman, 1987;

Menge et al., 1994), our study is the first to manipulate the hydrodynamic

environment to evaluate the possible deterrence of urchin attacks by wave‐induced

movement of algal fronds. Our wave tank trials demonstrated that the number of

urchins attacking a motionless D. viridis plant was orders of magnitude greater than

for plants moved about by waves. Moreover, the attack success of urchins on the alga

was inversely related to the amplitude of the wave‐induced sweeping motion of the

plants. This is consistent with our field observations that urchins attacks on D. viridis

are much greater during calmer conditions. Thus, both our laboratory and field

observations show that wave‐induced movements of D. viridis plays a major role in

protecting it from urchin grazing.

Although the repulsion of urchins appears to be due primarily to the indirect

effect of water motion causing sweeping of algal fronds, the urchins could also be

responding directly to water motion. This hypothesis is supported by the drop in the

number of urchins on the walls of the tank with increasing water motion. In the trials

with D. viridis, urchins were ~9 times more abundant on the walls at the end of the

trials with no sweep than in the trials with a 10‐cm sweep. Moreover, urchins did not

116

move after the first 20 min in the trials with the highest water motion. Current speeds

peaked at ~0.15 and 0.30 m s‐1 for 5‐cm and 10‐cm sweeping D. viridis, respectively

(as measured with a Doppler current meter), indicating that S. droebachiensis stops

moving at current speeds >0.30 m s‐1. This threshold current speed is less than half

than that reported in studies of effects of current speed on Strongylocentrotus nudus

(~0.70 m s‐1, Kawamata, 1998). Thus, water motion protects D. viridis from attacks by

S. droebachiensis first because hydrodynamic forces (e.g., drag, lift and acceleration

forces) decrease urchin movement and second due to the repulsion of urchins by the

sweeping of fronds. Given this hierarchical relationship, the deterrence of urchins by

the sweeping of algal fronds may only be important at intermediate flow speeds

when urchins are still moving.

The formation of dense aggregations appears to be an important part of the

foraging strategy of S. droebachiensis (Garnick, 1978; Bernstein et al., 1983; Vadas et al.,

1986), and can lead to the formation of grazing fronts than can advance through kelp

beds at a rate of up to 4.0 m month‐1. However, this only appears to occur when

urchins attain a critical biomass (Breen & Mann, 1976b; Scheibling et al., 1999;

Gagnon et al., in press). Our experiments demonstrate the key role of density in the

interaction between urchins and macroalgae. The relation of attack success to urchin

density is not linear but involves a threshold density, as with the formation of grazing

fronts. Our wave tank trials at different current speeds suggest that the critical

density increases with increased wave conditions.

In the Mingan Islands, D. viridis occurs exclusively on urchin barrens. Our field

observations indicate that urchins readily accumulate on, and consume D. viridis

under still conditions, but that urchins avoid this alga when the fronds move back

and forth with wave action. Thus, D. viridis is less well defended in the absence of

waves, and the chemical composition of D. viridis alone seems to provide little

protection from grazing. Why D. viridis produces acid remains unclear. Perhaps,

small amounts of acid are released when the fronds are swept across surfaces and act

synergistically with water motion to repel grazers, or the acid is more effective

117

against less voracious grazers (e.g., amphipods). The chemical defense may also only

be effective when more palatable algae are present. Alternatively, the primary role of

acid may not be a grazer deterrent, but rather act as an allelochemical or anti‐fouling

substance as shown for sulfuric acid and other chemicals in various species

interactions (Stoecker, 1978; Hay, 1996). Finally, D. viridis occurs in many regions

(Peters et al., 1997), and its capacity to produce acid might have evolved in response

to factors in other communities, and may not be a trait which provides advantages in

the northern Atlantic region.

Our study is the first to examine the interaction between chemical,

morphological and hydrodynamic factors in providing an alga with protection

against herbivory. We show that the survival of D. viridis in an intensively‐grazed

environment largely depends on its delicate structure. This contrasts with the

strategy of producing heavily calcified tissues as used by the coralline algae

commonly found on urchin barrens (Steneck, 1986). Beyond the documentation

of this novel plant defense, we have shown how the direct and indirect effects of

water motion can modulate species interactions, further providing evidence of the

importance of hydrodynamic factors in structuring of marine communities (Lissner,

1983; Witman & Dayton, 2001; Gagnon et al., 2003a). Finally, we demonstrate the

need to integrate a variety of biotic and abiotic variables, including algal chemistry

and morphology, water motion and temperature, and the abundance of predators, in

studies of algal defenses.

118

Appendix 5.1

Wave tank system. a) The rotational force generated by a 2‐HP electric motor is transmitted through a series of pillow blocks, axles, pulleys, and belts. (b) A perforated bar welded to the last axle (enlarged in the upper‐left corner, reverse view) is connected to the nearest of two plywood panels hinged to the bottom of the wave tank. Two metal bars connecting the top of the first panel with the top of the second causes the two panels to move back and forth in parallel producing oscillating waves.

Chapitre VI

Conclusion générale

120

Mon étude augmente notre connaissance des facteurs et mécanismes qui contrôlent

les communautés végétales de la zone infralittorale rocheuse de l’archipel de Mingan.

D’une part, elle constitue l’examen le plus détaillé de l’évolution temporelle de la

distribution et de l’abondance des macrophytes et quantifie pour la première fois

l’impact du broutage de l’oursin vert, Strongylocentrotus droebachiensis, sur la structure

et la régulation des assemblages algaux. D’autre part, elle identifie certains

mécanismes de défense qui permettent aux juvéniles d’Agarum cribrosum et à

Desmarestia viridis, les deux macrophytes dominants en zone dénudée, de résister au

broutage de l’oursin vert.

Dynamique de répartition des macrophytes

Le premier volet de l’étude démontre que la distribution et l’abondance des

macrophytes varient grandement selon (1) la communauté (zone des laminaires et

zone dénudée), (2) la période de l’année et (3) la localité (site) considérées. Les plus

grandes fluctuations surviennent dans la communauté des laminaires; pendant que

certaines ceintures reculent de plusieurs mètres en quelques mois, d’autres avancent

de façon tout aussi substantielle. Une large part de cette variabilité est attribuable au

broutage de l’oursin vert qui limite, de façon prononcée, le recrutement et la

succession algale, tel que démontré lors de l’exclusion d’oursins en zone dénudée. La

concordance entre la position des fronts d’oursins et la position de la limite inférieure

des ceintures tout au long de l’étude indique que l’effet du broutage se répercute à

l’échelle côtière en contrôlant la pénétration en profondeur des ceintures de

laminaires. Les retraits de ceintures les plus marqués surviennent pendant l’été,

lorsque les fronts de broutage s’intensifient, alors que la recolonisation a lieu

principalement en hiver, lorsque les populations d’oursins aux abords des ceintures

s’atténuent. L’efficacité d’un front de broutage à détruire une ceinture de laminaires

tend à augmenter avec la biomasse d’oursins dans le front. Cette efficacité serait

toutefois affectée par la composition végétale des ceintures. En effet, mes résultats

indiquent que la progression ascendante d’un front est retardée au contact d’agrégats

d’Agarum cribrosum et suggèrent fortement que l’établissement de larges bandes

121

d’A. cribrosum en marge inférieure des ceintures de laminaires augmente la stabilité

de ces dernières. Mes observations suggèrent que l’abrasion par les glaces en hiver

provoque de grandes fluctuations dans les populations d’oursins (par mortalité ou

dispersion des oursins).

En zone dénudée, la stabilité temporelle de la distribution et de l’abondance

des macrophytes charnus peut être étonnamment élevée malgré la forte abondance

d’oursins, comme c’est notamment le cas d’A. cribrosum. Les agrégats d’A. cribrosum

persistent pendant des périodes de plusieurs années (au moins deux ans), avec même

un léger taux d’accroissement. La résilience élevée des agrégats aux perturbations à

fine échelle (m) en fait un habitat relativement stable qui favorise le recrutement

algal, comme en témoigne l’abondance plus élevée de la rhodophycée Ptilota serrata et

des juvéniles d’A. cribrosum à l’intérieur qu’à l’extérieur des agrégats. Le recrutement

accru sous la canopée d’A. cribrosum s’expliquerait en partie par une réduction du

broutage de l’oursin dont l’abondance sous la canopée est faible (les oursins sont

majoritairement localisés à l’extérieur des agrégats). Les agrégats d’A. cribrosum

affectent donc positivement le reste de la communauté végétale en contribuant à

l’augmentation de la survie algale en zone dénudée.

Mécanismes de défense des macrophytes

Le second volet de l’étude démontre que la survie algale en zone dénudée repose

fondamentalement sur la modulation des interactions macrophytes‐oursins par les

propriétés hydrodynamiques du milieu. L’examen de la distribution des juvéniles

d’Agarum cribrosum en zone dénudée révèle qu’ils sont nettement plus nombreux sur

les aires balayées par les plants de Desmarestia viridis qu’en zone ouverte.

L’abondance des juvéniles sous un plant de D. viridis augmente avec la taille du plant

et décroît avec l’augmentation de la distance par rapport au centre du plant. Au

contraire, l’oursin est plus abondant à l’extérieur que sous un plant de D. viridis et son

abondance augmente avec la distance par rapport au centre du plant. Ce patron de

distribution serait principalement attribuable au mouvement de balayage de D. viridis

causé par les vagues, lequel mouvement empêcherait la pénétration des oursins sous

122

la canopée, résultant en une diminution du broutage de l’oursin sous la canopée et,

par conséquent, en un taux de survie accru des juvéniles. Les juvéniles sont

vulnérables au broutage de l’oursin; en moins de quatre jours, un juvénile dépourvu

d’un couvert de D. viridis subit une perte moyenne de plus de 50 % de sa longueur,

alors que cette perte n’est que de 3 % avec couvert. Même si la croissance des

juvéniles sous D. viridis est légèrement diminuée par rapport à leur croissance en

zone ouverte, cette perte est nettement compensée par le gain de protection contre le

broutage découlant de l’association. Mes résultats indiquent donc que la défense par

association avec D. viridis augmente considérablement la survie des juvéniles

d’A. cribrosum. Cette association défensive éphémère (D. viridis disparaît à l’automne)

est proposée comme étape de succession menant à la formation d’agrégats adulte

d’A. cribrosum.

L’investigation, en milieu contrôlé, de la réponse comportementale de l’oursin

face à D. viridis révèle que la propriété répulsive de l’algue est davantage liée à

l’interaction entre ses attributs morphologiques et l’action des vagues, qu’à sa

composition chimique. En l’absence de vagues, les oursins s’agrègent rapidement sur

l’algue immobile, alors qu’ils l’évitent en eau agitée, lorsque l’algue balaie le substrat.

L’intensité des attaques, telle que déterminée par le nombre d’oursins en contact avec

l’algue, diminue avec l’augmentation de la force des vagues et augmente avec

l’abondance des oursins. Le déplacement de l’oursin et l’intensité de ses attaques

contre D. viridis semblent augmenter avec l’accroissement de la température. L’oursin

n’est repoussé par l’odeur de D. viridis qu’à des températures correspondant aux

extrêmes estivaux dans l’archipel de Mingan. Ces résultats, couplés aux observations

in situ d’agrégats d’oursins s’alimentant aisément sur D. viridis en eau calme,

indiquent que la composition chimique de l’algue (e.g., acide sulfurique) ne contribue

que faiblement à la défense de l’algue contre le broutage de l’oursin.

Portée de l’étude

Mon étude contribue de plusieurs façons à l’avancement de notre connaissance de

l’écologie des communautés végétales marines et des interactions plantes‐herbivores:

123

1) Elle identifie l’oursin vert, Strongylocentrotus droebachiensis, comme facteur‐clé dans

la régulation de la distribution et de l’abondance des macrophytes dans l’archipel

de Mingan et fournit la première preuve quantitative que d’importantes

fluctuations saisonnières surviennent dans les populations d’oursins et de

macrophytes, tant au niveau de leur abondance que de leur structure de taille;

2) Elle est la seule étude ayant quantifié la dynamique de répartition naturelle des

assemblages algaux infralittoraux dans le golfe du Saint‐Laurent (et l’une des rares

dans le nord‐ouest de l’Atlantique). Ce genre d’étude est essentiel à notre

compréhension du fonctionnement de cet écosystème dans lequel les assemblages

algaux sont d’une grande importance pour le reste de la communauté;

3) Elle identifie les interactions positives (e.g., refuge par association) comme

mécanisme important pour la survie algale infralittorale et fournit, du coup, un

argument de poids aux écologistes de plus en plus nombreux à reconnaître l’impact

significatif de telles interactions dans la structure et la régulation des

communautés;

4) Elle souligne l’importance d’examiner plusieurs facteurs biotiques (e.g., attributs

morphologiques et chimiques des plantes et abondance de l’herbivore) et de

mesurer l’impact de variations dans les paramètres physiques (e.g., température,

conditions hydrodynamiques) dans les études d’interactions plantes‐herbivores;

5) Elle fournit la première preuve expérimentale quantitative que l’hydrodynamisme

peut moduler les interactions plantes‐herbivores en milieu marin en exerçant à la

fois un effet direct et indirect sur la quête alimentaire d’un herbivore.

La plupart des modèles élaborés pour tenter d’expliquer la structure des

communautés ont accordé une grande importance aux interactions négatives telles

que la prédation et la compétition (e.g., zones intertidale et infralittorale). Cet intérêt

scientifique élevé pour les interactions négatives a relégué en arrière‐plan les

interactions positives dont la reconnaissance de l’importance, en tant que mécanisme

structurant les communautés, s’est accrue au cours des dernières années

124

(e.g., marais salés). Les interactions positives, telles que le mutualisme et le

commensalisme, s’opèrent par l’entremise d’espèces fondatrices dont la présence

améliore la qualité d’un habitat (pour un ou plusieurs autres organismes de la

communauté) en modifiant ses propriétés physico‐chimiques ou en augmentant

l’hétérogénéité de l’environnement, créant ainsi des refuges spatiaux contre les stress

environnementaux ou la prédation. Dans un environnement où la prédation est

élevée et le recrutement des proies est faible, les interactions positives peuvent jouer

un rôle important dans la structure des communautés et contribuer à augmenter la

diversité spécifique.

Mon étude s’inscrit bien dans ce récent mouvement de reconnaissance du rôle

communautaire important que peuvent jouer les interactions positives. En effet, mes

résultats indiquent que l’oursin vert exerce une prédation considérable sur les

communautés algales infralittorales de l’archipel de Mingan, résultant en une baisse

marquée de l’abondance de plusieurs espèces (e.g., laminaires). Néanmoins, les

formes adultes (résistantes au broutage de l’oursin) des phaéophycées

Agarum cribrosum et Desmarestia viridis contribuent au maintien de la diversité

végétale en zone dénudée. Les agrégats d’A. cribrosum et les plaques de D. viridis

augmentent l’hétérogénéité spatiale de la zone dénudée et forment des micro‐habitats

répulsifs à l’oursin vert et à l’intérieur desquels le recrutement algal (e.g., juvéniles

d’A. cribrosum et Ptilota serrata) est nettement plus élevé qu’en milieu ouvert. À ce

titre, A. cribrosum et D. viridis semblent représenter deux espèces fondatrices, l’une

(A. cribrosum) exerçant un effet positif sur une période potentielle de plusieurs années

(espèce pérenne), l’autre (D. viridis) sur une période de quelques mois (espèce

annuelle). Peu étudiées, les interactions algales positives, telles que décrites ici dans

l’archipel de Mingan (e.g., refuge par association), pourraient s’avérer d’une

importance sous‐estimée dans la structure des communautés infralittorales marines.

Nouvelles questions et hypothèses

Mes travaux ont soulevé d’intéressantes questions et généré de nouvelles hypothèses

qui méritent un examen attentif:

125

1) Les patrons de variation observés dans la distribution et l’abondance des

macrophytes sont‐ils représentatifs des patrons à plus grande échelle? La réponse à

cette question passe impérativement par le recours à une méthode plus

performante que la plongée sous‐marine utilisée dans le cadre de cette étude et ne

pouvant suffire à quantifier de tels patrons que sur de courtes distances

(e.g, une centaine de mètres) au prix d’investissements considérables en temps et

efforts de plongée. L’utilisation de la cartographie infralittorale par imagerie

aérienne, bien que plus onéreuse que la plongée, s’avère extrêmement prometteuse

(des percées majeures dans ce domaine sont en cours) et se justifie par la réduction

substantielle du temps et des efforts d’échantillonnage ainsi que par l’élargissement

considérable des échelles spatiales et temporelles (i.e., sur des centaines de

kilomètres et pendant des décennies). La distribution en faible profondeur de la

plupart des macrophytes et la clarté habituelle de l’eau dans le nord du golfe du

Saint‐Laurent favorisent l’utilisation des techniques de photographies aériennes;

2) L’abrasion hivernale et printanière par les glaces exerce un impact notable sur la

structure des communautés intertidales dans le golfe et l’estuaire du Saint‐Laurent.

Mes résultats indiquent que l’effet des glaces dans l’archipel de Mingan s’étend à

des profondeurs pouvant dépasser 5 m et que l’établissement de la laminaire

Agarum cribrosum survient principalement en hiver. Il serait souhaitable de mener

d’autres études afin de tester rigoureusement les deux hypothèses selon lesquelles

l’abrasion par les glaces (1) contrôle les populations d’oursins en causant la

mortalité de ces derniers ou en provoquant leur migration en zones plus profondes

et (2) facilite l’établissement d’A. cribrosum en zone dénudée;

3) La présence de Desmarestia viridis en zone dénudée augmente nettement la survie

des juvéniles d’A. cribrosum en procurant à ces derniers un refuge éphémère contre

le broutage de l’oursin. Une investigation sur une plus longue période que celle

impliquée dans cette étude (i.e., >5 mois), incluant le suivi de juvéniles avant et

après la période hivernale, est requise afin d’évaluer l’importance de ce refuge par

association dans la formation d’agrégats adultes d’A. cribrosum;

126

4) Mes résultats indiquent que le taux de retrait des ceintures de laminaires tend à

augmenter avec la biomasse d’oursins dans les fronts de broutage. Il semble

toutefois que cette relation ne s’applique pas à l’ensemble des sites étudiés et

qu’elle puisse même changer de façon saisonnière. D’autres suivis temporels de la

position de la limite inférieure des ceintures de laminaires et de la structure des

fronts d’oursins, incluant une taille échantillonnale accrue et des mesures

automnales et hivernales, sont requis afin de mieux caractériser le comportement

de cette relation;

5) L’exploitation commerciale des stocks naturels d’oursins dans l’archipel de Mingan

est‐elle possible et si oui, comment procéder? Mes résultats indiquent qu’une telle

pratique devrait cibler les fronts de broutage aux abords des ceintures de

laminaires où l’abondance et la taille des oursins sont supérieures

comparativement à tout autre endroit du littoral. Ceci favoriserait la récolte rapide

des individus les plus susceptibles de rencontrer les standards commerciaux

(e.g., taille du test et index gonado‐somatique). D’autre part, l’élevage de stocks

d’oursins en mer (sea ranching) pourrait s’effectuer par l’adition d’oursins à

l’intérieur d’ouvertures pratiquées dans de larges agrégats d’A. cribrosum.

Les oursins seraient alors contenus par cette barrière algale naturelle, nourris par

l’apport anthropique de macrophytes et retirés à l’approche de l’hiver afin de

prévenir d’éventuelles pertes d’oursins causées par l’abrasion par les glaces.

127

Bibliographie

Ackerman J.D. & Okubo A. (1993). Reduced mixing in a marine macrophyte canopy. Funct. Ecol. 7: 305‐309.

Anderson R.J. & Velimirov B. (1982). An experimental investigation of the palatability of kelp bed algae to the sea urchin Parechinus angulosus Leske. Mar. Ecol. 3: 357‐373.

Archambault D. & Bourget E. (1983). Importance du régime de dénudation sur la structure et la succession des communautés intertidales de substrat rocheux en milieu subarctique. Can. J. Fish. Aquat. Sci. 40: 1278‐1292.

Atsatt P.R. & OʹDowd D.J. (1976). Plant defense guilds. Many plants are functionally interdependent with respect to their herbivores. Science 193: 24‐29.

Balch T. & Scheibling R.E. (2000). Temporal and spatial variability in settlement and recruitment of echinoderms in kelp beds and barrens in Nova Scotia. Mar. Ecol. Prog. Ser. 205: 139‐154.

Bazhin A.G. (1998). The sea urchin genus Strongylocentrotus in the seas of Russia. In: Mooi R. & Telford M. (Eds.), Echinoderms: San Francisco. Balkema, Rotterdam, pp. 563‐566.

Becerra J.X., Venable D.L., Evans P.H. & Bowers W.S. (2001). Interactions between chemical and mechanical defenses in the plant genus Bursera and their implications for herbivores. Amer. Zool. 41: 865‐876.

Bégin C. (2002). Influence de la couverture algale sur les communautés dʹinvertébrés benthiques dans le nord du Golfe du St‐Laurent. M.Sc. Thesis, Université Laval, Québec.

Begon M., Harper J.L. & Townsend C.R. (1996). Ecology. Blackwell Science, Oxford. Benedetti‐Cecchi L. (2000). Predicting direct and indirect interactions during

succession in a mid‐littoral rocky shore assemblage. Ecol. Monogr. 70: 45‐72. Bergeron P. & Bourget E. (1984). Effet du froid et des glaces sur les peuplements

intertidaux des régions nordiques particulièrement dans lʹestuaire du Saint‐Laurent. Oceanis 10: 279‐304.

Bernstein B.B., Williams B.E. & Mann K.H. (1981). The role of behavioral responses to predators in modifying urchinsʹ (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.

Bertness M.D. & Callaway R. (1994). Positive interactions in communities. Trends Ecol. Evol. 9: 191‐193.

Bologna P.A.X. & Steneck R.S. (1993). Kelp beds as habitats for American lobster Homarus americanus. Mar. Ecol. Prog. Ser. 100: 127‐134.

128

Bolser R.C. & Hay M.E. (1996). Are tropical plants better defended? Palatability and defenses of temperate vs. tropical seaweeds. Ecology 77: 2269‐2286.

Bourget E., Archambault D. & Bergeron P. (1985). Effet des propriétés hivernales sur les peuplements épibenthiques intertidaux dans un milieu subarctique, lʹestuaire du Saint‐Laurent. Nat. Can. 112: 131‐142.

Branch G.M. (1975). Mechanisms reducing intraspecific competition in Patella spp.: migration, differentiation and territorial behaviour. J. Anim. Ecol. 44: 575‐600.

Brawley S.H. (1992). Mesoherbivores. In: John D.M., Hawkins S.J. & Price J.H. (Eds.), Plant‐animal interactions in the marine benthos. Oxford University Press, Oxford, pp. 235‐263.

Breen P.A. & Mann K.H. (1976a). Changing lobster abundance and the destruction of kelp beds by sea urchins. Mar. Biol. 34: 137‐142.

Breen P.A. & Mann K.H. (1976b). Destructive grazing of kelp by sea urchins in eastern Canada. J. Fish. Res. Bd. Can. 33: 1278‐1283.

Briscoe C.S. & Sebens K.P. (1988). Omnivory in Strongylocentrotus droebachiensis (Müller) (Echinodermata: Echinoidea): predation on subtidal mussels. J. Exp. Mar. Biol. Ecol. 115: 1‐24.

Brown B.J. & Ewel J.J. (1987). Herbivory in complex and simple tropical successional ecosystems. Ecology 68: 108‐116.

Brown V.K. & Gange A.C. (1992). Secondary plant succession: how is it modified by insect herbivory? Vegetatio 101: 3‐13.

Bruno J.F. & Bertness M.D. (2001). Habitat modification and facilitation in benthic marine communities. In: Bertness M.D., Gaines S.D. & Hay M.E. (Eds.), Marine community ecology. Sinauer Associates, Sunderland, pp. 201‐218.

Bulleri F., Benedetti‐Cecchi L. & Cinelli F. (1999). Grazing by the sea urchins Arbacia lixula L. and Paracentrotus lividus Lam. in the Northwest Mediterranean. J. Exp. Mar. Biol. Ecol. 241: 81‐95.

Callaway R.M. (1995). Positive interactions among plants. Bot. Rev. 61: 306‐349. Callaway R.M., Kikvidze Z. & Kikodze D. (2000). Facilitation by unpalatable weeds

may conserve plant diversity in overgrazed meadows in the Caucasus Mountains. Oikos 89: 275‐282.

Carpenter R.C. (1986). Partitioning herbivory and its effects on coral reef algal communities. Ecol. Monogr. 56: 345‐363.

Castilla J.C. & Crisp D.J. (1970). Responses of Asterias rubens to olfactory stimuli. J. Mar. Biol. Assoc. UK. 50: 829‐847.

Cetrulo G.L. & Hay M.E. (2000). Activated chemical defenses in tropical versus temperate seaweeds. Mar. Ecol. Prog. Ser. 207: 243‐253.

Chapman A.R.O. (1981). Stability of sea urchin dominated barren grounds following destructive grazing of kelp in St. Margaretʹs Bay, eastern Canada. Mar. Biol. 62: 307‐311.

129

Chapman A.R.O. & Johnson C.R. (1990). Disturbance and organization of macroalgal assemblages in the Northwest Atlantic. Hydrobiologia 192: 77‐121.

Chapman M.G. & Underwood A.J. (1992). Foraging behaviour of marine benthic grazers. In: John D.M., Hawkins S.J. & Price J.H. (Eds.), Plant‐animal interactions in the marine benthos. Oxford University Press, New York, pp. 289‐317.

Charnov E.L. (1976). Optimal foraging, the marginal value theorem. Theor. Popul. Biol. 9: 129‐136.

Coley P.D. & Barone J.A. (1996). Herbivory and plant defenses in tropical forests. Annu. Rev. Ecol. Syst. 27: 305‐335.

Connell J.H. (1961). The influence of interspecific competition and other factors on the distribution of the barnacle Chthamalus stellatus. Ecology 42: 710‐723.

Connell J.H. & Slatyer R.O. (1977). Mechanisms of succession in natural communities and their role in community stability and organization. Am. Nat. 111: 1119‐1144.

Conover W.J. (1980). Practical Nonparametric Statistics. John Wiley & Sons, New York.

Cowen R.K., Agegian C.R. & Foster M.S. (1982). The maintenance of community structure in a central California giant kelp forest. J. Exp. Mar. Biol. Ecol. 64: 189‐201.

Crowder M.J. & Hand D.J. (1990). Analysis of repeated measures. Chapman & Hall, London.

Davidson D.W. (1993). The effects of herbivory and granivory on terrestrial plant succession. Oikos 68: 23‐35.

Dayton P.K. (1985a). The structure and regulation of some South American kelp communities. Ecol. Monogr. 55: 447‐468.

Dayton P.K. (1985b). Ecology of kelp communities. Annu. Rev. Ecol. Syst. 16: 215‐245. Dayton P.K., Tegner M.J., Parnell P.E. & Edwards P.B. (1992). Temporal and spatial

patterns of disturbance and recovery in a kelp forest community. Ecol. Monogr. 62: 421‐445.

Dayton P.K., Currie V., Gerrodette T., Keller B.D., Rosenthal R. & Ven Tresca D. (1984). Patch dynamics and stability of some California kelp communities. Ecol. Monogr. 54: 253‐289.

Dean T.A. & Jacobsen F.R. (1984). Growth of juvenile Macrocystis pyrifera (Laminariales) in relation to environmental factors. Mar. Biol. 83: 301‐311.

Dean T.A., Schroeter S.C. & Dixon J.D. (1984). Effects of grazing by two species of sea urchins (Strongylocentrotus franciscanus and Lytechinus anamesus) on recruitment and survival of two species of kelp (Macrocystis pyrifera and Pterygophora californica). Mar. Biol. 78: 301‐313.

130

Draper N.R. & Smith H. (1981). Applied regression analysis. John Wiley & Sons, New York.

Duffy J.E. & Hay M.E. (1990). Seaweed adaptations to herbivory. BioScience 40: 368‐375.

Duffy J.E. & Hay M.E. (2000). Strong impacts of grazing amphipods on the organization of a benthic community. Ecol. Monogr. 70: 237‐263.

Duffy J.E. & Hay M.E. (2001). The ecology and evolution of marine consumer‐prey interactions. In: Bertness M.D., Gaines S.D. & Hay M.E. (Eds.), Marine community ecology. Sinauer Associates, Sunderland, pp. 131‐157.

Duggins D.O. (1980). Kelp beds and sea otters: an experimental approach. Ecology 61: 447‐453.

Duggins D.O. (1981). Sea urchins and kelp: the effects of short term changes in urchin diet. Limnol. Oceanogr. 26: 391‐394.

Duggins D.O., Eckman J.E. & Sewell A.T. (1990). Ecology of understory kelp environments. II. Effects of kelps on recruitment of benthic invertebrates. J. Exp. Mar. Biol. Ecol. 143: 27‐45.

Ebeling A.W., Laur D.R. & Rowley R.J. (1985). Severe storm disturbances and reversal of community structure in a southern California kelp forest. Mar. Biol. 84: 287‐294.

Eckman J.E., Duggins D.O. & Sewell A.T. (1989). Ecology of understory kelp environments. I. Effects of kelps on flow and particle transport near the bottom. J. Exp. Mar. Biol. Ecol. 129: 173‐187.

Emlen J.M. (1966). The role of time and energy in food preference. Am. Nat. 100: 611‐617.

Eppley R.W. & Bovell C.R. (1958). Sulfuric acid in Desmarestia. Biol. Bull. 115: 101‐106. Estes J.A., Smith N.S. & Palmisano J.F. (1978). Sea otter predation and community

organization in the western Aleutian Islands, Alaska. Ecology 59: 822‐833. Falk‐Petersen I.‐B. & Lonning S. (1983). Reproductive cycles of two closely related sea

urchin species, Strongylocentrotus droebachiensis (O. F. Müller) and Strongylocentrotus pallidus (G. O. Sars). Sarsia 68: 157‐164.

Forseth I.N., Wait D.A. & Casper B.B. (2001). Shading by shrubs in a desert system reduces the physiological and demographic performance of an associated herbaceous perennial. J. Ecol. 89: 670‐680.

Gagnon P., Wagner G. & Himmelman J.H. (2003a). 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.

Gagnon P., Himmelman J.H. & Johnson L.E. (2003b). Algal colonization in urchin barrens: defense by association during recruitment of the brown alga Agarum cribrosum. J. Exp. Mar. Biol. Ecol. 290: 179‐196.

131

Gagnon P., Himmelman J.H. & Johnson L.E. (in press). Temporal variation in community interfaces: kelp bed boundary dynamics adjacent to persistent urchin barrens. Mar. Biol.

Garnick E. (1978). Behavioral ecology of Strongylocentrotus droebachiensis (Müller) (Echinodermata: Echinoidea). Aggregating behavior and chemotaxis. Oecologia 37: 77‐84.

Gaymer C.F., Himmelman J.H. & Johnson L.E. (2001). Distribution and feeding ecology of the seastars Leptasterias polaris and Asterias vulgaris in the northern Gulf of St. Lawrence, Canada. J. Mar. Biol. Assoc. UK. 81: 827‐843.

Gerard V.A. (1982). In situ water motion and nutrient uptake by the giant kelp Macrocystis pyrifera. Mar. Biol. 69: 51‐54.

Gerard V.A. (1984). The light environment in a giant kelp forest: influence of Macrocystis pyrifera on spatial and temporal variability. Mar. Biol. 84: 189‐195.

Graham L.E. & Wilcox L.W. (2000). Algae. Prentice‐Hall, Upper Saddle River. Gray J.S. (1977). The stability of benthic ecosystems. Helgoländer wiss. Meeresunters.

30: 427‐444. Hagen N.T. (1983). Destructive grazing of kelp beds by sea urchins in Vestfjorden,

northern Norway. Sarsia 68: 177‐190. Hand D.J. & Taylor C.C. (1987). Multivariate analysis of variance and repeated

measures. Chapman & Hall, London. Harrold C. & Reed D.C. (1985). Food availability, sea urchin grazing, and kelp forest

community structure. Ecology 66: 1160‐1169. Hay M.E. (1986). Associational plant defenses and the maintenance of species

diversity: turning competitors into accomplices. Am. Nat. 128: 617‐641. Hay M.E. (1996). Marine chemical ecology: whatʹs known and whatʹs next? J. Exp.

Mar. Biol. Ecol. 200: 103‐134. Hay M.E. & Fenical W. (1988). Marine plant‐herbivore interactions: the ecology of

chemical defense. Annu. Rev. Ecol. Syst. 19: 111‐145. Hay M.E. & Fenical W. (1992). Chemical mediation of seaweed‐herbivore interactions.

In: John D.M., Hawkins S.J. & Price J.H. (Eds.), Plant‐animal interactions in the marine benthos. Oxford University Press, New York, pp. 319‐337.

Hay M.E., Kappel Q.E. & Fenical W. (1994). Synergisms in plant defenses against herbivores: interactions of chemistry, calcification, and plant quality. Ecology 75: 1714‐1726.

Hendrix S.D., Brown V.K. & Gange A.C. (1988). Effects of insect herbivory on early plant succession: comparison of an English site and an American site. Biol. J. Linn. Soc. 35: 205‐216.

Himmelman J.H. (1969). Some aspects of the ecology of Strongylocentrotus droebachiensis in eastern Newfoundland. M.Sc. Thesis, Memorial University of Newfoundland, St. Johnʹs.

132

Himmelman J.H. (1975). Phytoplankton as a stimulus for spawning in three marine invertebrates. J. Exp. Mar. Biol. Ecol. 20: 199‐214.

Himmelman J.H. (1984). Urchin feeding and macroalgal distribution in Newfoundland, eastern Canada. Nat. Can. 111: 337‐348.

Himmelman J.H. (1986). Population biology of green sea urchins on rocky barrens. Mar. Ecol. Prog. Ser. 33: 295‐306.

Himmelman J.H. & Steele D.H. (1971). Foods and predators of the green sea urchin Strongylocentrotus droebachiensis in Newfoundland waters. Mar. Biol. 9: 315‐322.

Himmelman J.H. & Lavergne Y. (1985). Organization of rocky subtidal communities in the St. Lawrence Estuary. Nat. Can. 112: 143‐154.

Himmelman J.H. & Nédélec 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., 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.

Holling C.S. (1973). Resilience and stability of ecological systems. Annu. Rev. Ecol. Syst. 4: 1‐23.

Hooper R. (1981). Recovery of Newfoundland benthic marine communities from sea ice. Proc. Int. Seaweed Symp. 8: 360‐366.

Hurd C.L., Stevens C.L., Laval B.E., Lawrence G.A. & Harrison P.J. (1997). Visualization of seawater flow around morphologically distinct forms of the giant kelp Macrocystis integrifolia from wave‐sheltered and exposed sites. Limnol. Oceanogr. 42: 156‐163.

Jensen M. (1974). The Strongylocentrotidae (Echinoidea), a morphologic and and systematic study. Sarsia 57: 113‐148.

Johnson C.R. & Mann K.H., 1982. Adaptations of Strongylocentrotus droebachiensis for survival on barren grounds in Nova Scotia. In: Lawrence J.M. (Ed.), International Echinoderms Conference. Balkema, A. A., Tampa Bay, pp. 277‐283.

Johnson C.R. & Mann K.H. (1988). Diversity, patterns of adaptation, and stability of Nova Scotian kelp beds. Ecol. Monogr. 58: 129‐154.

Jones N.S. & Kain J.M. (1967). Subtidal algal colonization following the removal of Echinus. Helgoländer wiss. Meeresunters. 15: 460‐466.

Karban R. & Myers J.H. (1989). Induced plant responses to herbivory. Annu. Rev. Ecol. Syst. 20: 331‐348.

Kawamata S. (1998). Effect of wave‐induced oscillatory flow on grazing by a subtidal sea urchin Strongylocentrotus nudus (A. Agassiz). J. Exp. Mar. Biol. Ecol. 224: 31‐48.

133

Kawamata S. (2001). Adaptive mechanical tolerance and dislodgement velocity of the kelp Laminaria japonica in wave‐induced water motion. Mar. Ecol. Prog. Ser. 211: 89‐104.

Keats D.W., South G.R. & Steele D.H. (1982). The occurence of Agarum cribrosum (Mert.) Bory (Phaeophyta, Laminariales) in relation to some of its competitors and predators in Newfoundland. Phycologia 21: 189‐191.

Keats D.W., South G.R. & Steele D.H. (1985). Algal biomass and diversity in the upper subtidal at a pack‐ice disturbed site in eastern Newfoundland. Mar. Ecol. Prog. Ser. 25: 151‐158.

Keats D.W., South G.R. & Steele D.H. (1990). Effects of an experimental reduction in grazing by green sea urchins on a benthic macroalgal community in eastern Newfoundland. Mar. Ecol. Prog. Ser. 68: 181‐193.

Kerr J.N.Q. & Paul V.J. (1995). Animal‐plant defense association: the soft coral Sinularia sp. (Cnidaria, Alcyonacea) protects Halimeda spp. from herbivory. J. Exp. Mar. Biol. Ecol. 186: 183‐205.

Konar B. (2000). Seasonal inhibitory effects of marine plants on sea urchins: structuring communities the algal way. Oecologia 125: 208‐217.

Krebs C.J. (2001). Ecology. Benjamin Cummings, San Francisco. Larson B.R., Vadas R.L. & Keser M. (1980). Feeding and nutritional ecology of the sea

urchin Strongylocentrotus drobachiensis in Maine, USA. Mar. Biol. 59: 49‐62. Lawrence J.M. (1975). On the relationships between marine plants and sea urchins.

Oceanogr. Mar. Biol. Ann. Rev. 13: 213‐286. Legendre P. & Legendre L. (1998). Numerical ecology. Elsevier, Amsterdam. Leinaas H.P. & Christie H. (1996). Effects of removing sea urchins (Strongylocentrotus

droebachiensis): stability of the barren state and succession of kelp forest recovery in the east Atlantic. Oecologia 105: 524‐536.

Lettenmaier D.P. (1976). Detection of trends in water quality data from records with dependent observations. Water Resour. Res. 12: 1037‐1046.

Lewis S.M. (1986). The role of herbivorous fishes in the organization of a Caribbean reef community. Ecol. Monogr. 56: 183‐200.

Lewontin R.C. (1969). The meaning of stability. Brookhaven Symp. Biol. 22: 13‐24. Lindroth R.L. (1989). Mammalian herbivore‐plant interactions. In: Abrahamson W.G.

(Ed.), Plant‐animal interactions. McGraw‐Hill, New York, pp. 163‐206. Lissner A.L. (1983). Relationship of water motion to the shallow water distribution

and morphology of two species of sea urchins. J. Mar. Res. 41: 691‐709. Littler M.M., Littler D.S. & Taylor P.R. (1987). Animal‐plant defense associations:

effects on the distribution and abundance of tropical reef macrophytes. J. Exp. Mar. Biol. Ecol. 105: 107‐121.

Lobban C.S. & Harrison P.J. (1994). Seaweed ecology and physiology. Cambridge University Press, Cambridge.

134

Lubchenco J. & Gaines S.D. (1981). A unified approach to marine plant‐herbivore interactions. I. Populations and communities. Annu. Rev. Ecol. Syst. 12: 405‐437.

Lucas P.W., Turner I.M., Dominy N.J. & Yamashita N. (2000). Mechanical defences to herbivory. Ann. Bot.‐London. 86: 913‐920.

MacArthur R.H. & Pianka E.R. (1966). On optimal use of a patchy environment. Am. Nat. 100: 603‐609.

Mann K.H. (1973). Seaweeds: their productivity and strategy for growth. Science 182: 975‐981.

Mann K.H. (1977). Destruction of kelp‐beds by sea‐urchins: A cyclical phenomenon or irreversible degradation? Helgol. wiss. Meeresunters. 30: 455‐467.

Mann K.H., Wright J.L.C., Welsford B.E. & Hatfield E. (1984). Responses of the sea urchin Strongylocentrotus droebachiensis (O. F. Müller) to water‐borne stimuli from potential predators and potential food algae. J. Exp. Mar. Biol. Ecol. 79: 233‐244.

McAuliffe J.R. (1986). Herbivore‐limited establishment of a Sonoran Desert tree, Cercidium microphyllum. Ecology 67: 276‐280.

McClintock M., Higinbotham N., Uribe E.G. & Cleland R.E. (1982). Active, irreversible accumulation of extreme levels of H2SO4 in the brown alga, Desmarestia. Plant. Physiol. 70: 771‐774.

McCullagh P. & Nelder J.A. (1989). Generalized linear models. Chapman & Hall, New York.

McNaughton S.J. (1983). Compensatory plant growth as a response to herbivory. Oikos 40: 329‐336.

Meidel S.K. & Scheibling R.E. (1998). Annual reproductive cycle of the green sea urchin, Strongylocentrotus droebachiensis, in differing habitats in Nova Scotia, Canada. Mar. Biol. 131: 461‐478.

Meidel S.K. & Scheibling R.E. (1999). Effects of food type and ration on reproductive maturation and growth of the sea urchin Strongylocentrotus droebachiensis. Mar. Biol. 134: 155‐166.

Menge B.A. (1972). Foraging strategy of a starfish in relation to actual prey availability and environmental predictability. Ecol. Monogr. 42: 25‐50.

Menge B.A. (1976). Organization of the New England rocky intertidal community: role of predation, competition, and environmental heterogeneity. Ecol. Monogr. 46: 355‐391.

Menge B.A., Berlow E.L., Blanchette C.A., Navarrete S.A. & Yamada S.B. (1994). The keystone species concept: variation in interaction strength in a rocky intertidal habitat. Ecol. Monogr. 64: 249‐286.

Miller R.J. (1985). Succession in sea urchin and seaweed abundance in Nova Scotia, Canada. Mar. Biol. 84: 275‐286.

135

Miller R.J. & Colodey A.G. (1983). Widespread mass mortalities of the green sea urchin in Nova Scotia, Canada. Mar. Biol. 73: 263‐267.

Minor M.A. & Scheibling R.E. (1997). Effects of food ration and feeding regime on growth and reproduction of the sea urchin Strongylocentrotus droebachiensis. Mar. Biol. 129: 159‐167.

Montgomery D.C. (1991). Design and analysis of experiments. John Wiley & Sons, New York.

Morrison D. (1988). Comparing fish and urchin grazing in shallow and deeper coral reef algal communities. Ecology 69: 1367‐1382.

Mortensen T.H. (1943). A monograph of the Echinoidea. Reitzels, C. A., Copenhagen. Myers J.H. & Bazely D. (1991). Thorns, spines, prickles, and hairs: are they stimulated

by herbivory and do they deter herbivores? In: Tallamy D.W. & Raupp M.J. (Eds.), Phytochemical induction by herbivores. John Wiley, New York, pp. 325‐344.

Nestler E.C. & Harris L.G. (1994). The importance of omnivory in Strongylocentrotus droebachiensis (Muller) in the Gulf of Maine. In: David B., Guille A., Féral J.‐P. & Roux M. (Eds.), Echinoderms through time. Balkema, Rotterdam, pp. 813‐818.

Novaczek I. & McLachlan J. (1989). Investigations of the marine algae of Nova Scotia XVII: vertical and geographic distribution of marine algae on rocky shores of the Maritime Provinces. Proc. N. S. Inst. Sci. 38: 91‐143.

OʹDowd D.J. & Williamson G.B. (1979). Stability conditions in plant defense guilds. Am. Nat. 114: 379‐383.

Odum E.P. (1969). The strategy of ecosystem development. Science 164: 262‐270. Ogden J.C. & Lobel P.S. (1978). The role of herbivorous fishes and urchins in coral

reef communities. Env. Biol. Fish. 3: 49‐63. Paine R.T. (1966). Food web complexity and species diversity. Am. Nat. 100: 65‐75. Paine R.T. (1976). Size‐limited predation: an observational and experimental

approach with the Mytilus‐Pisaster interaction. Ecology 57: 858‐873. Paine R.T. (1990). Benthic macroalgal competition: complications and consequences. J.

Phycol. 26: 12‐17. Paine R.T. & Vadas R.L. (1969). The effects of grazing by sea urchins,

Strongylocentrotus spp., on benthic algal populations. Limnol. Oceanogr. 14: 710‐719.

Paul V.J. & Hay M.E. (1986). Seaweed susceptibility to herbivory: chemical and morphological correlates. Mar. Ecol. Prog. Ser. 33: 255‐264.

Paul V.J. & Van Alstyne K.L. (1992). Activation of chemical defenses in the tropical green alga Halimeda spp. J. Exp. Mar. Biol. Ecol. 160: 191‐203.

Pavia H. & Toth G.B. (2000). Inducible chemical resistance to herbivory in the brown seaweed Ascophyllum nodosum. Ecology 81: 3212‐3225.

136

Pearse J.S. & Hines A.H. (1979). Expansion of a central California kelp forest following the mass mortality of sea urchins. Mar. Biol. 51: 83‐91.

Pelletreau K.N. & Muller‐Parker G. (2002). Sulfuric acid in the phaeophyte alga Desmarestia munda deters feeding by the sea urchin Strongylocentrotus droebachiensis. Mar. Biol. 141: 1‐9.

Peres‐Neto P.R. (1999). How many statistical tests are too many? The problem of conducting multiple ecological inferences revisited. Mar. Ecol. Prog. Ser. 176: 303‐306.

Peters A.F., van Oppen M.J.H., Wiencke C., Stam W.T. & Olsen J.L. (1997). Phylogeny and historical ecology of the Desmarestiaceae (Phaeophyceae) support a southern hemisphere origin. J. Phycol. 33: 294‐309.

Pfister C.A. & Hay M.E. (1988). Associational plant refuges: convergent patterns in marine and terrestrial communities result from differing mechanisms. Oecologia 77: 118‐129.

Prince J.S. & LeBlanc W.G. (1992). Comparative feeding preference of Strongylocentrotus droebachiensis (Echinoidea) for the invasive seaweed Codium fragile ssp. tomentosoides (Chlorophyceae) and four other seaweeds. Mar. Biol. 113: 159‐163.

Randall J.E. (1961). Overgrazing of algae by herbivorous marine fishes. Ecology 42: 812.

Raymond B.G. & Scheibling R.E. (1987). Recruitment and growth of the sea urchin Strongylocentrotus droebachiensis (Muller) following mass mortalities off Nova Scotia, Canada. J. Exp. Mar. Biol. Ecol. 108: 31‐54.

Reed D.C. (1990). The effects of variable settlement and early competition on patterns of kelp recruitment. Ecology 71: 776‐787.

Reed D.C. & Foster M.S. (1984). The effects of canopy shading on algal recruitment and growth in a giant kelp forest. Ecology 65: 937‐948.

Rice W.R. (1989). Analyzing tables of statistical tests. Evolution. 43: 223‐225. Rochette R., Hamel J.‐F. & Himmelman J.H. (1994). Foraging strategy of the asteroid

Leptasterias polaris: role of prey odors, current and feeding status. Mar. Ecol. Prog. Ser. 106: 93‐100.

Rose C.D., Sharp W.C., Kenworthy W.J., Hunt J.H., Lyons W.G., Prager E.J., Valentine J.F., Hall M.O., Whitfield P.E. & Fourqurean J.W. (1999). Overgrazing of a large seagrass bed by the sea urchin Lytechinus variegatus in Outer Florida Bay. Mar. Ecol. Prog. Ser. 190: 211‐222.

Rosenthal J.P. & Kotanen P.M. (1994). Terrestrial plant tolerance to herbivory. Trends Ecol. Evol. 9: 145‐148.

Ruiz G.M., Carlton J.T., Grosholz E.D. & Hines A.H. (1997). Global invasions of marine and estuarine habitats by non‐indigenous species: mechanisms, extent, and consequences. Amer. Zool. 37: 621‐632.

137

Russell M.P., Ebert T.A. & Petraitis P.S. (1998). Field estimates of growth and mortality of the green sea urchin, Strongylocentrotus droebachiensis. Ophelia 48: 137‐153.

SAS Institute Inc. (1999). SAS/STAT Userʹs Guide, Version 8. SAS Institute Inc, Cary, NC.

Sasaki H., Kataoka H., Kamiya M. & Kawai H. (1999). Accumulation of sulfuric acid in Dictyotales (Phaeophyceae): taxonomic distribution and ion chromatography of cell extracts. J. Phycol. 35: 732‐739.

Schappert P.J. & Shore J.S. (1999). Cyanogenesis, herbivory and plant defense in Turnera ulmifolia on Jamaica. Écoscience 6: 511‐520.

Scheibling R. (1986). Increased macroalgal abundance following mass mortalities of sea urchins (Strongylocentrotus droebachiensis) along the Atlantic coast of Nova Scotia. Oecologia 68: 186‐198.

Scheibling R.E. (1988). Microbial control of sea urchins: Achillesʹ heel or Pandoraʹs box? In: Burke R.D., Mladenov P.V., Lambert P. & Parsley R.L. (Eds.), Echinoderm biology. Balkema, Rotterdam, pp. 745‐754.

Scheibling R.E. (1996). The role of predation in regulating sea urchin populations in eastern Canada. Oceanol. Acta 19: 421‐430.

Scheibling R.E. & Stephenson R.L. (1984). Mass mortality of Strongylocentrotus droebachiensis (Echinodermata: Echinoidea) off Nova Scotia, Canada. Mar. Biol. 78: 153‐164.

Scheibling R.E. & Raymond B.G. (1990). Community dynamics on a subtidal cobble bed following mass mortalities of sea urchins. Mar. Ecol. Prog. Ser. 63: 127‐145.

Scheibling R.E. & Hamm J. (1991). Interactions between sea urchin (Strongylocentrotus droebachiensis) and their predators in field and laboratory experiments. Mar. Biol. 110: 105‐116.

Scheibling R.E. & Hennigar A.W. (1997). Recurrent outbreaks of disease in sea urchins Strongylocentrotus droebachiensis in Nova Scotia: evidence for a link with large‐scale meteorologic and oceanographic events. Mar. Ecol. Prog. Ser. 152: 155‐165.

Scheibling R.E. & Hatcher B.G. (2001). The ecology of Strongylocentrotus droebachiensis. In: Lawrence J.M. (Ed.), Edible sea urchins: biology and ecology. Elsevier, Amsterdam, pp. 271‐306.

Scheibling R.E., Hennigar A.W. & Balch T. (1999). Destructive grazing, epiphytism, and disease: the dynamics of sea urchin‐kelp interactions in Nova Scotia. Can. J. Fish. Aquat. Sci. 56: 2300‐2314.

Schmitz O.J., Hambäck P.A. & Beckerman A.P. (2000). Trophic cascades in terrestrial systems: a review of the effects of carnivore removals on plants. Am. Nat. 155: 141‐153.

138

Schnitzler I., Pohnert G., Hay M. & Boland W. (2001). Chemical defense of brown algae (Dictyopteris spp.) against the herbivorous amphipod Ampithoe longimana. Oecologia 126: 515‐521.

Schoener T.W. (1974). Resource partitioning in ecological communities. Science 185: 27‐39.

Sebens K.P. (1985). Community ecology of vertical rock walls in the Gulf of Maine, U.S.A.: small‐scale processes and alternative community states. In: Moore P.G. & Seed R. (Eds.), The ecology of rocky coasts. Hodder and Stoughton, London, pp. 346‐371.

Seymour R.J., Tegner M.J., Dayton P.K. & Parnell P.E. (1989). Storm wave induced mortality of giant kelp, Macrocystis pyrifera, in Southern California. Estuar. Coast. Shelf S. 28: 277‐292.

Siddon C.E. & Witman J.D. (2003). Influence of chronic, low‐level hydrodynamic forces on subtidal community structure. Mar. Ecol. Prog. Ser. 261: 99‐110.

Sih A., Crowley P., McPeek M., Petranka J. & Strohmeier K. (1985). Predation, competition and prey communities: a review of field experiments. Annu. Rev. Ecol. Syst. 16: 269‐311.

Simenstad C.A., Estes J.A. & Kenyon K.W. (1978). Aleuts, sea otters, and alternate stable‐state communities. Science 200: 403‐411.

Sivertsen K. (1997). Geographic and environmental factors affecting the distribution of kelp beds and barren grounds and changes in biota associated with kelp reduction at sites along Norwegian coast. Can. J. Fish. Aquat. Sci. 54: 2872‐2887.

Sloan N.A. & Campbell A.C. (1982). Perception of food. In: Jangoux M. & Lawrence J.M. (Eds.), Echinoderm nutrition. Balkema, A. A., Rotterdam, pp. 3‐23.

Snedecor G.W. & Cochran W.G. (1989). Statistical methods. Iowa State University Press, Ames.

Sousa W.P. (1984). The role of disturbance in natural communities. Annu. Rev. Ecol. Syst. 15: 353‐391.

Sousa W.P. & Connell J.H. (1992). Grazing and succession in marine algae. In: John D.M., Hawkins S.J. & Price J.H. (Eds.), Plant‐animal interactions in the marine benthos. Oxford University Press, New York, pp. 425‐441.

Starr M., Himmelman J.H. & Therriault J.‐C. (1990). Direct coupling of marine invertebrate spawning with phytoplankton blooms. Science 247: 1071‐1074.

Steneck R.S. (1983). Escalating herbivory and resulting adaptive trends in calcareous algal crusts. Paleobiology 9: 44‐61.

Steneck R.S. (1985). Adaptations of crustose coralline algae to herbivory: patterns in space and time. In: Toomey D.F. & Nitecki M.H. (Eds.), Paleoalgology: contemporary research and applications. Springer‐Verlag, Heidelberg, pp. 352‐366.

139

Steneck R.S. (1986). The ecology of coralline algal crusts: convergent patterns and adaptative strategies. Annu. Rev. Ecol. Syst. 17: 273‐303.

Stoecker D. (1978). Resistance of a tunicate to fouling. Biol. Bull. 155: 615‐626. Sutherland J.P. (1974). Multiple stable points in natural communities. Am. Nat. 108:

859‐873. Swan E.F. (1961). Some observations on the growth rate of sea urchins in the genus

Strongylocentrotus. Biol. Bull. 120: 420‐427. Tahvanainen J.O. & Root R.B. (1972). The influence of vegetational diversity on the

population ecology of a specialized herbivore, Phyllotreta cruciferae (Coleoptera:Chrysomelidae). Oecologia 10: 321‐346.

Taylor R.B., Sotka E. & Hay M.E. (2002). Tissue‐specific induction of herbivore resistance: seaweed response to amphipod grazing. Oecologia 132: 68‐76.

Tegner M.J. & Dayton P.K. (2000). Ecosystem effects of fishing in kelp forest communities. ICES J. Mar. Sci. 57: 579‐589.

Thompson R.J. (1984). Partitioning of energy between growth and reproduction in three populations of the green sea urchin Strongylocentrotus droebachiensis. Adv. Invertebr. Reprod. 3: 425‐432.

Thompson T.E. (1988). Acidic allomones in marine organisms. J. Mar. Biol. Assoc. UK. 68: 499‐517.

Tiffin P. (2000). Mechanisms of tolerance to herbivore damage: what do we know? Evol. Ecol. 14: 523‐536.

Tremblay C. & Chapman A.R.O. (1980). The local occurrence of Agarum cribrosum in relation to the presence or absence of its competitors and predators. Proc. N. S. Inst. Sci. 30: 165‐170.

Trowbridge C.D. (1995). Establishment of the green alga Codium fragile ssp. tomentosoides on New Zealand rocky shores: current distribution and invertebrate grazers. J. Ecol. 83: 949‐965.

Underwood A.J. (1999). Physical disturbances and their direct effect on an indirect effect: responses of an intertidal assemblage to a severe storm. J. Exp. Mar. Biol. Ecol. 232: 125‐140.

Underwood A.J. & Jernakoff P. (1981). Effects of interactions between algae and grazing gastropods on the structure of a low‐shore intertidal algal community. Oecologia 48: 221‐233.

Vadas R.L. (1968). The ecology of Agarum and the kelp bed community. Ph.D. Thesis, University of Washington, Seattle.

Vadas R.L. (1977). Preferential feeding: an optimization strategy in sea urchins. Ecol. Monogr. 47: 337‐371.

Vadas R.L. (1990). Comparative foraging behavior of tropical and boreal sea urchins. In: Hughes R.N. (Ed.), Behavioural mechanisms of food selection. Springer‐Verlag, Berlin, pp. 479‐514.

140

Vadas R.L. & Elner R.W. (1992). Plant‐animal interactions in the north‐west Atlantic. In: John D.M., Hawkins S.J. & Price J.H. (Eds.), Plant‐animal interactions in the marine benthos. Oxford University Press, Oxford, pp. 33‐60.

Vadas R.L., Wright W.A. & Miller S.L. (1990). Recruitment of Ascophyllum nodosum: wave action as a source of mortality. Mar. Ecol. Prog. Ser. 61: 263‐272.

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.

Van Alstyne K.L. (1988). Herbivore grazing increases polyphenolic defenses in the intertidal brown alga Fucus distichus. Ecology 69: 655‐663.

Van Alstyne K.L., Ehlig J.M. & Whitman S.L. (1999). Feeding preferences for juvenile and adult algae depend on algal stage and herbivore species. Mar. Ecol. Prog. Ser. 180: 179‐185.

Vasquez J.A. (1993). Effects on the animal community of dislodgment of holdfasts of Macrocystis pyrifera. Pac. Science 47: 180‐184.

Vasquez J.A. & Buschmann A.H. (1997). Herbivore‐kelp interactions in Chilean subtidal communities: a review. Rev. Chil. Hist. Nat. 70: 41‐52.

Vasquez J.A. & McPeak R.H. (1998). A new tool for kelp restoration. Calif. Fish Game 84: 149‐158.

Velimirov B. & Griffiths C.L. (1979). Wave‐induced kelp movement and its importance for community structure. Bot. Mar. 22: 169‐172.

Walker L.R., Zasada J.C. & Chapin III F.S. (1986). The role of life history processes in primary succession on an Alaskan floodplain. Ecology 67: 1243‐1253.

Weis A.E. & Berenbaum M.R. (1989). Herbivorous insects and green plants. In: Abrahamson W.G. (Ed.), Plant‐animal interactions. McGraw‐Hill, New York, pp. 123‐162.

White P.S. (1979). Pattern, process, and natural disturbance in vegetation. Bot. Rev. 45: 229‐299.

Witman J.D. (1985). Refuges, biological disturbance, and rocky subtidal community structure in New England. Ecol. Monogr. 55: 421‐445.

Witman J.D. (1987). Subtidal coexistence: storms, grazing, mutualism and the zonation of kelps and mussels. Ecol. Monogr. 57: 167‐187.

Witman J.D. & Dayton P.K. (2001). Rocky subtidal communities. In: Bertness M.D., Gaines S.D. & Hay M.E. (Eds.), Marine community ecology. Sinauer Associates, Sunderland, pp. 339‐366.

Wood D.M. & del Moral R. (1987). Mechanisms of early primary succession in subalpine habitats on Mount St. Helens. Ecology 68: 780‐790.

Zamora R., Hodar J.A. & Gomez J.M. (1999). Plant‐herbivore interaction: beyond a binary vision. In: Pugnaire F.I. & Valladares F. (Eds.), Handbook of functional plant ecology. Marcel Dekker, Inc., New York, pp. 677‐718.

141

Zar J.H. (1999). Biostatistical analysis. Prentice‐Hall, New Jersey.

Documents

Dynamique de répartition et mécanismes de défense des ... · III Abstract In the northern Gulf of St. Lawrence, the green sea urchin, Strongylocentrotus droebachiensis, is the