Vol. 370: 45–52, 2008 MARINE ECOLOGY PROGRESS SERIES Published October 28 doi: 10.3354/meps07568 Mar Ecol Prog Ser Behavioral mechanism for an associational refuge for macroalgae on temperate reefs Stuart Levenbach* Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, California 93106, USA Present address: Office of Management and Budget, New Executive Office Bldg., 725 17th Street NW, Room 9215, Washington, DC 20503, USA ABSTRACT: Temperate reefs are sensitive to fluctuations in grazing intensity and, while some spe- cies of marine algae possess chemical and structural defenses, many species lack adaptations to with- stand high grazer densities. The presence of foliose macroalgae in the face of high grazing pressure can have important consequences for higher trophic levels, in part because macroalgae harbor prey for fish and macroinvertebrates. The present study explores a biotic interaction between the sea urchins Strongylocentrotus purpuratus and S. franciscanus and the colonial anemone Corynactis cal- ifornica, and describes the behavioral underpinnings of an associational refuge for benthic macroal- gae on heavily grazed rocky reefs. Field surveys of heavily grazed reefs revealed that macroalgae is more abundant within C. californica colonies, and that sea urchins touch far fewer C. californica polyps than would be expected by chance. In laboratory studies, the movement of urchins was impeded when urchins withdrew their tube feet upon contact with C. californica. This was probably the behavioral mechanism explaining results of field experiments where recruitment of macroalgae was relatively high among C. californica colonies transplanted within a sea urchin barren. C. cali- fornica may play an underappreciated role on temperate rocky reefs by altering the behavior of a dominant herbivore and thus providing an associational refuge for important biogenic structure. KEY WORDS: Facilitation · Associational defense · Sea urchin · Corynactis californica · Macroalgae · Tube feet Resale or republication not permitted without written consent of the publisher INTRODUCTION 2003), and deterring herbivores with noxious chemi- cals (Callaway et al. 2000, Paul et al. 2001) or stinging Herbivory is a dominant force affecting the distribu- tentacles (Littler et al. 1987). tion and abundance of terrestrial and marine plants, The most recent research on associational plant and many species possess chemical and structural defenses has focused on heavily grazed terrestrial traits that deter herbivores (Rosenthal & Kotanen 1994, plant communities (Milchunas & Noy-Meir 2002, Re- McClintock & Baker 2001). For species lacking these bollo et al. 2002, Baraza et al. 2006), yet some of the adaptations, studies of positive interactions have shed earliest examples came from marine ecosystems (Hay light on the importance of associational defenses, 1986, Pfister & Hay 1988). In marine ecosystems, tem- whereby one species gains protection when in close perate reefs are frequently denuded of canopy forming proximity to another facilitator species (Stachowicz kelp and understory macroalgae by sea urchins (Law- 2001). Facilitators have been shown to reduce herbi- rence 1975, Harrold & Reed 1985, Watanabe & Harrold vory on vulnerable neighbors in a variety of ways, such 1991), with extant foliose macroalgae confined to as concealing target species from detection (Atsatt & ledges and other physical refuges that are inaccessible O’Dowd 1976, Hamback et al. 2000), impeding access to dominant grazers. Associational defenses that en- (Kerr & Paul 1995, Rebollo et al. 2002, Gagnon et al. able benthic macroalgae to persist may have important *Email: [email protected] © Inter-Research 2008 · www.int-res.com 46 Mar Ecol Prog Ser 370: 45–52, 2008 ramifications for temperate reef communities because National Park Service’s Kelp Forest Monitoring Pro- macroalgae harbor amphipods and other crustaceans gram were used to examine the distributions of Cory- that are a critical source of food for benthic fish (Laur & nactis and sea urchins at the among-reef scale, which Ebeling 1983, Holbrook & Schmitt 1992). included 15 rocky reefs on 5 islands (Santa Barbara, The present study focuses on the behavioral mecha- Anacapa, Santa Cruz, Santa Rosa, and San Miguel). nism underlying an associational refuge on temperate Data used for this analysis were from 2001, the year reefs involving the small colonial anemone Corynactis that had the largest range in the percent cover of Cory- californica (hereafter Corynactis). This species has nactis (Kushner et al. 2004), and hence the greatest particularly powerful nematocysts (Skaer & Picken chance of detecting an antagonistic relationship be- 1965), aggressively uses its mesenterial filaments to tween Corynactis and sea urchins. On each reef, den- kill neighboring sessile organisms (Chadwick 1987), sities of red and purple urchins were counted in twenty and can impede movement of sea stars (Patton et al. 2 m2 quadrats placed at random intervals along a 100 × 1991). The following questions were addressed in the 4 m transect at depths ranging from 7 to 15 m. The per- present study: (1) Does the distribution of Corynactis cent of reef surface covered by Corynactis was esti- explain significant variation in the abundance of sea mated by identifying the primary substrate holder and urchins on reefs where they co-occur?; (2) Do sea any overlapping taxa at 40 randomly spaced points urchins avoid contact with Corynactis and does their within a 0.5 × 3 m area at 25 random locations along behavioral response vary among species of urchins? the transect (for details, see Davis 1988). Simple linear (3) Can Corynactis facilitate the recruitment of benthic regressions were used to determine the extent to macroalgae within urchin barrens? I hypothesized that which the cover of Corynactis predicted the density of if sea urchins are impeded by the stinging tentacles of purple and red urchins. Corynactis, then sea urchins should be less common in The percent cover of Corynactis and density of ur- areas where Corynactis is dense, exhibit signs of stress chins were surveyed in 0.25 m2 patches on horizontally when in contact with Corynactis, and avoid contact oriented bedrock at Naples Reef, an isolated outcrop with polyps. If the above hypotheses are true, then located 2 km offshore and 23 km west of Santa Bar- Corynactis could facilitate recruitment of macroalgae bara, CA (34° 25’ N, 119° 57’ W; Ebeling et al. 1985). amidst high densities of sea urchins on barren reefs. The density of red and purple urchins and cover of Corynactis on Naples Reef was measured in sixty-nine 50 × 50 cm quadrats placed approximately 2 m apart at MATERIALS AND METHODS a water depth of 9 to 12 m in June 2003. Quadrats were gridded with monofilament line and the primary space Distribution of macroalgae among Corynactis. Car- holder was identified at 81 uniformly spaced points pinteria Reef (Santa Barbara Channel, CA, USA; within each quadrat. The extent to which cover of 34° 23’ N, 119° 32’ W) is a shallow (<12 m) rocky reef Corynactis predicted the density of purple and red heavily grazed by red (Strongylocentrotus francis- urchins was assessed using simple linear regression. canus) and purple urchins (S. purpuratus). The reef was To examine the distribution of Corynactis at the scale surveyed to determine whether macroalgae were more of individual urchins, the number of Corynactis polyps abundant among Corynactis colonies than adjacent ar- touching each urchin was compared with what would eas without Corynactis. Nine Corynactis patches (all be expected if urchins were randomly distributed over oriented horizontally) were surveyed in July 2005, each the same area. Two non-adjacent and horizontally ori- between 20 and 50 cm in diameter at a water depth of ented areas on Naples Reef (2 × 10 m) were surveyed 7 m. An 18 × 9 cm frame was placed in the center and in February 2002 (N = 717 purple urchins, 39 red ur- then outside the perimeter of each Corynactis patch. chins for both areas combined). Kelp (e.g. Macrocystis The area within each frame was photographed and the pyrifera and Pterygophora californica) was rare and percent cover of benthic taxa was estimated by super- most urchins were considered to be foraging, based on imposing a grid of at least 30 points over each photo- being in open, exposed microhabitat. For each urchin graph and recording the primary space holder. in an exposed microhabitat, the test diameter was Spatial distribution of sea urchins and Corynactis. recorded with a ruler and also the number of Cory- To determine whether sea urchins were less common nactis polyps that the urchin touched. To test whether in areas where Corynactis is dense, the distribution of urchins were touching fewer polyps than expected, a Corynactis was compared with that of sea urchins at 3 purple urchin (5 cm test diameter; N = 692) and a red spatial scales: (1) in 400 m2 areas among reefs sepa- urchin (9 cm; N = 279) were placed on the reef and the rated by 10 to 50 km; (2) in 0.25 m2 areas within a sin- number of Corynactis polyps in contact with the urchin gle reef; and (3) within the area occupied by an indi- were counted. A chi-square test was used to compare vidual urchin. Data obtained from the Channel Islands the number of Corynactis polyps in contact with ran- Levenbach: Impediments to herbivory by sea urchins 47 domly placed purple and red urchins to the actual formed to determine whether Corynactis was capable number of polyps observed touching those that were of impeding foraging urchins. Both red and purple naturally distributed. urchins were collected from pier pilings adjacent to Effect of Corynactis on urchin behavior. To test UCSB-MSI and starved for 2 wk in flow-through sea- whether urchins avoided contact with Corynactis, I water tanks at UCSB-MSI in December 2003. Mussel placed red and purple urchins on substrata with or shells, covered with Corynactis, were manipulated to without Corynactis and recorded movement and tube create 3 levels of cover: 0, 30, and 100%.