Mar Biol (2008) 155:273–280 DOI 10.1007/s00227-008-1023-y

ORIGINAL PAPER

Macroalgal morphology mediates particle capture by the corallimorpharian californica

Kathleen M. Morrow · Robert C. Carpenter

Received: 21 June 2007 / Accepted: 18 June 2008 / Published online: 10 July 2008 © Springer-Verlag 2008

Abstract The shallow forest at Santa Catalina E. arborea canopy did not aVect particle capture. However, Island, California (33.45 N, ¡118.49 W) is distinguished D. undulata, found predominantly outside of the canopy, by several canopy guilds ranging from a Xoating canopy inhibited particle capture rates by 40% by redirecting parti- (), to a stipitate, erect understory can- cles around C. californica polyps and causing contraction opy ( arborea), to a short prostrate canopy just of the feeding . These results suggest that the mor- above the substratum (Dictyopteris, Gelidium, Laminaria, phology of Xexible marine organisms may aVect the distri- Plocamium spp.), followed by algal turfs and encrusting bution and abundance of adjacent passive suspension coralline algae. The prostrate macroalgae found beneath feeders. E. arborea canopies are primarily branching red algae, while those in open habitats are foliose . Densi- ties of , are signiWcantly greater Introduction under E. arborea canopies than outside (approximately 1,200 versus 300 polyps m¡2 respectively). Morphological Benthic suspension feeders occupy the majority of avail- diVerences in macroalgae between these habitats may aVect able hard substratum in subtidal rocky reef environments, the rate of C. californica particle capture and serve as a which results in frequent competitive and facilitative inter- mechanism for determining distribution and abun- actions among neighboring organisms for access to space, dance. Laboratory experiments in a unidirectional Xume light, and/or food particles (Jackson 1977; Knowlton and under low (9.5 cm s¡1) and high (21 cm s¡1) Xow speeds Jackson 2001). Community structure of sessile biota is examined the eVect of two morphologically distinct macro- mediated in part, by the strength of these interspeciWc inter- algae on the capture rate of Artemia sp. cysts by C. califor- actions. We examine here how diVerences between mor- nica polyps. These experiments (January–March 2006) phologically distinct macroalgal taxa can interact with tested the hypothesis that a foliose macroalga, D. undulata, water Xow to mediate the particle capture abilities of adja- would inhibit particle capture more than a branching alga, cent suspension feeders. G. robustum. G. robustum, found predominantly under the Although cnidarians co-exist with macroalgae on a wide variety of substrata in tropical and temperate marine habi- tats, few studies have demonstrated the competitive or Communicated by J.P. Grassle. facilitative mechanisms by which these two groups interact (Jompa and McCook 2002). Algal morphology and Xexibil- & K. M. Morrow ( ) ity has important implications for cohabitants through the Department of Biological Sciences, Auburn University, 101 Rouse Life Sciences Building, Auburn, AL 36849, USA alteration of water Xow, which can modify the rate of sedi- e-mail: [email protected] mentation, gas exchange, and food delivery (Eckman et al. 1989; Hurd and Stevens 1997; Irving and Connell 2006). R. C. Carpenter Previous mechanistic studies have focused primarily on Department of Biology, California State University Northridge, 18111 NordhoV St., Northridge, CA 91330-8303, USA symbiotic hermatypic and mechanisms of algal over- e-mail: [email protected] growth. In the present study, Corynactis californica does 123 274 Mar Biol (2008) 155:273–280 not have photosynthesizing symbionts and relies exclu- exposed habitats where E. arborea is found (Charters et al. sively on heterotrophic feeding. Therefore, experiments 1968). High wave energy constantly moves the fronds and allow us to test speciWcally the mechanisms by which reduces shading and sedimentation beneath the canopy, algae inXuence feeding success when uninhibited by algal while scouring the substratum and supplying inorganic shading. nutrients to understory algae and benthic invertebrates Water Xow determines the delivery rate of food particles (Kennelly 1989; Eckman and Duggins 1991; Connell to suspension feeders (Sebens et al. 1997). The potential 2003a; Connell 2003b). food available for capture is determined by the Xux of parti- Our previous studies suggested that the presence of mac- cles past the feeding surface, which is a function of particle roalgae has a positive eVect on the abundance of C. califor- concentration and the rate at which particles are encoun- nica, both underneath the E. arborea canopy and in open tered by the feeding surface (Okamura and Partridge 1999). habitats (Morrow and Carpenter 2008). However, a signiW- If all other variables are held constant, an increase in Xow cantly greater density of C. californica polyps (fourfold will increase the number of particles exposed to the feeding diVerence) and a lower density of macroalgae are found surface. However, the delivery of particles is disrupted by under E. arborea canopies. Predominant macroalgal taxa roughness elements such as the shape, size, Xexibility and under the E. arborea canopy were upright, branching red proximity of associated organisms and bottom topography species, while, foliose and decumbent brown species were (McFadden 1986; Koehl 1996; Sebens et al. 1997). The found outside the canopy (Morrow and Carpenter 2008). functional form and density of macroalgae can inXuence The present study investigated whether morphologically macro- and microhabitat water Xow in benthic marine envi- distinct macroalgae selected from under and outside the E. ronments, thus mediating food delivery to adjacent suspen- arborea canopy could inXuence the rate of particle capture sion feeders (Anderson and Charters 1982; Johnson 2001). by C. californica and thereby alter their distribution and In the present study C. californica (the strawberry anem- abundance between habitats. one) was used as a model suspension feeder to examine the We hypothesized that the branching red alga, Gelidium eVect of two distinct macroalgae on the rate of particle cap- robustum, would inhibit particle capture to a lesser extent ture. These anemones form large aggregations of small than the foliose brown alga, Dictyopteris undulata, due to polyps (<15 mm diameter) on rocky substrata in kelp forest diVerences in morphology and Xexibility. G. robustum and communities oV the northeastern PaciWc from British D. undulata were chosen because they represent the most Columbia to Baja, California (Chadwick 1987). They abundant species found under and outside the E. arborea reproduce asexually by longitudinal Wssion and sexually canopy, respectively. We also hypothesized that rates of C. with a seasonal cycle of gametogenesis, producing plank- californica particle capture would be lower at high Xow tonic larvae from November through January (Holts and (21 cm s¡1) than at low Xow speeds (9.5 cm s¡1), due to Beauchamp 1993). Compared to scleractinian and actiniar- deformation of the polyp tentacles and inXuence of Xexible ian , C. californica has a rapid rate of asexual macroalgal thalli as drag increased (Koehl 1977). reproduction, doubling clonal populations in <2 mo and Gelidium robustum has a single holdfast and thalli that reaching densities of t2,400 polyps m¡2 on horizontal sub- are pinnately branched with an irregular pattern of closely strata, interspersed with macroalgae, at Santa Catalina spaced cylindrical branches which are stiV and cartilagi- Island, California (Chadwick 1987; Chadwick and Adams nous (Graham and Wilcox 2000). Although G. robustum 1991; Morrow and Carpenter 2008). Large asexual aggre- fronds are closely spaced, the pattern is suYciently open to gations of unknown size may be distinguished by color allow the Xow of water through the thalli (Anderson and morph (pink, brown, orange, and white), although the basis Charters 1982). The mesh-like structure and narrow branch for this color variation is unknown (Holts and Beauchamp diameter of species similar to G. robustum slow Xow and 1993). The polyps are consistently spaced »8 mm apart on dampen turbulence at Xow speeds <12 cm s¡1, transitioning horizontal surfaces at Santa Catalina Island. The distance to turbulent Xow at speeds >14 cm s¡1 (Anderson and between nearest neighbors was measured using vernier cal- Charters 1982; Hurd and Stevens 1997). Turbulent Xow can ipers in haphazardly placed quadrats (Morrow 2006). create recirculating eddies behind algal thalli and may Shallow, wave-exposed environments at Santa Catalina lengthen the time that particles spend in the vicinity of Island are distinguished by stipitate such as Eisenia polyp tentacles, thereby increasing capture rates. D. undu- arborea, as well as diverse assemblages of foliose and lata is a foliose brown alga with thalli that are Xattened and geniculate coralline macroalgae. E. arborea occurs highly branched in a repeatedly dichotomous pattern and throughout temperate regions of the PaciWc Ocean and can attached to the substratum by a small holdfast (Graham and strongly inXuence abiotic factors (e.g. light and water Xow) Wilcox 2000). Flexible algal thalli, as in D. undulata, can aVecting the community structure of understory species. be deformed passively into streamlined shapes at high Xow Maximum water velocities can reach 14 m s¡1 in shallow, speeds, a trait that reduces drag and dislodgement, but may 123 Mar Biol (2008) 155:273–280 275 cause abrasion and retraction of nearby polyps as thalli con- tact the polyp surface (Koehl 1996). Finally, the eVective- ness of the contrast between these two algal morphologies relies on the assumption that D. undulata and G. robustum have similar biological eVects (e.g., allelochemicals, dis- solved organics) on C. californica polyps. We assumed that any chemicals emitted by the algal thalli would be diluted in the large volume of water (26 l) used during experimen- tal trials. Fig. 1 Diagram of working section in 26-l recirculating experimental Xume. A 2.5 £ 2.5-cm tile with Wve attached polyps was placed 50 cm downstream from Xow straighteners. Flow direction was from right to left, passing hydrated Artemia sp. cysts over three algal thalli that were Materials and methods 5 cm upstream of the polyps

Specimens of C. californica were collected using SCUBA from dense aggregations of polyps on a vertical rock wall at C. californica polyps. These cysts have been used exten- Bird Rock, Santa Catalina Island, California (33.45N, sively as food particles in feeding experiments with cnidari- ¡118.49W). To minimize sampling polyps from a single ans (Sebens 1981; Patterson 1991), although there has been clone, clusters of polyps attached to fragments of rock were some discussion about the merits of using passive particles removed along a 50-m transect at 1-m intervals at 2–5 m such as Artemia sp. cysts (Patterson 1991; Trager et al. depth. Polyps were transported in seawater to recirculating 1994). seawater tables at California State University Northridge Prior to each experiment, 0.5 ml of dried cysts were within 24 h of collection,. Individual polyps with pedal disc hydrated for ¸2h in Wltered seawater and then slowly diameters of 3.0 § 0.1 mm were removed from the substra- added »10 cm upstream of C. californica polyps. Cyst con- tum so that there remained a small piece of rock or coralline centrations in the Xume were quantiWed by siphoning 10-ml algae attached to the pedal disc. Polyps were attached to samples from 10 cm downstream of polyps and at a level 25 £ 25-mm clear acrylic tiles with cyanoacrylate glue comparable to the height of the C. californica tentacles. (Loctite® gel) in groups of Wve, with one polyp on each Samples of cysts (n = 3) were taken at random throughout corner and one in the center. Polyps were arranged at natural each experiment and counted using a dissecting microscope densities for horizontal substrata and spaced »8mm (Helmuth and Sebens 1993). Values were averaged to apart. Tiles with attached polyps were maintained in obtain the mean number of cysts available throughout each aerated aquaria at ambient seawater temperatures (16–18C) experiment to determine whether cyst concentrations Xuc- and fed Artemia sp. once weekly. tuated within or between experimental trials. Feeding by All experiments were conducted from January–March the polyps never resulted in detectable depletion of cysts 2006 in a 26-l laboratory Xume (working Sect. 100-cm during experiments (Fig. 2). long, 10-cm wide, 10-cm deep), similar to that described by Vogel and LaBarbera, 1978 (Fig. 1). A unidirectional Xow regime was created within the Xume using a motor- driven propeller (Dayton 1/90 HP). Experiments were run at 9.5 (low Xow) and 21 cm s¡1 (high Xow). Two sets of Xow straighteners were mounted at the entrance of the Xume channel to maintain a constant velocity proWle across the tank. The speed control settings and Xow veloc- ities in the working section of the Xume were calibrated by tracking the distance a hydrated neutrally buoyant Artemia sp. cyst traveled in a known time period based on camera shutter speed and using Scion Image software. A rectangu- lar piece of Plexiglas designed to hold the tiles with attached polyps was mounted with silicone sealant approximately 50-cm downstream of the Xow straighten- Fig. 2 Mean particle capture (§SE) polyp¡1 over 5 min in each algal ers (Fig. 1). X X X  treatment at two ow speeds. Because neither ow speed nor ow Hydrated Artemia sp. cysts (average diameter 250 m), speed/algal treatment had signiWcant eVects in a 3-factor ANOVA were used as experimental food particles because they are (Table 1), data for the two Xow speeds were combined (n = 20 for neutrally buoyant, proteinaceous, and readily captured by No-Algae treatment and n = 22 for algal treatments) 123 276 Mar Biol (2008) 155:273–280

A factorial statistical design was used to assess the eVects dardized for each tile as the total number of cysts captured, of; (1) algal presence/absence, (2) algal species (branching divided by the number of expanded polyps at the beginning vs. foliose), and (3) Xow speed (9.5 and 21 cm s¡1) on parti- of each experiment (particles captured polyp¡1). cle capture rates of C. californica. Flow velocities quantiWed Ten C. californica tiles were assigned randomly to each over one year at Bird Rock, Santa Catalina Island with an S4 treatment (No-Algae, Branching red algae, and Foliose current meter 0.5 m above the substratum at 3 m depth brown algae) and run at both Xow speeds (i.e., each repli- (where C. californica and macroalgae are found), ranged cate tile was used twice, once at each Xow speed). Each tile from 6 to 70 cm s¡1 (Carpenter unpubl. data) so the experi- recovered from transport and feeding for 48 h before it was mental speeds selected for this experiment lay within this used at the second Xow speed. To test the hypothesis that range. The eVects of algal presence/absence and species algal morphology aVected particle capture rate at either morphology were tested using three treatments: (1) No- Xow speed, data were analyzed using a mixed-model Algae (control), (2) G. robustum (branching red alga) and, ANOVA with the following factors: algal treatment (n =3), (3) D. undulata (foliose brown alga). For each treatment, Xow speed (n = 2), treatment by Xow speed interaction, three algal thalli were cut to 6-cm lengths that extended ver- and replicate tiles nested within treatment (Table 1). Algal tically (5 cm) into freestream Xow, once secured within the treatment interactions for each Xow speed were analyzed Xume. The three algal thalli were oriented perpendicular to with individual 1-way ANOVA’s and Tukey post-hoc the Xow and the holdfasts were threaded through a plastic tests using adjusted alpha values ( = 0.017). The per- base and secured to the Xume bottom with inert putty centage of C. californica polyps that retracted was calculated (Sculpey II®). The base was secured 4 cm upstream of the for each treatment, to determine whether Xow speed or tile, allowing the 5-cm tall thalli to cover the C. californica algal species aVected retraction. The number of polyps if deformed downstream by water Xow (Fig. 1). open polyps at the end of a feeding trial was divided by the The Xume was Wlled with artiWcial seawater and number open at the beginning of the trial and multiplied by enclosed by a water jacket chilled to ambient seawater tem- 100 to yield the percent of polyps that remained open after peratures (16–18 C) to maintain the working section at a each trial. To determine whether Xow speed or algal treat- constant temperature. Hydrated Artemia sp. cysts were ment impacted polyp retraction rates, arcsin transformed added to the Xume and allowed to circulate. Tiles with C. percentages were analyzed with a two-way ANOVA and californica polyps were transferred from the holding Tukey post-hoc tests. Finally, the longest and shortest aquaria and placed in chilled, Wltered seawater for at least diameters of each pedal disc, as well as the height of each 1 h prior to feeding experiments to allow digestion of previ- retracted C. californica polyp, were measured to the nearest ously ingested food. Prior analysis of C. californica gut contents determined that full digestion of adult Artemia sp. Table 1 (a) C. californica particle capture rates analyzed using a within a polyp occurs in 30–45 min. At the beginning of 3-factor ANOVA among algal treatments (No-Algae, G. robustum, each experiment, cysts were blocked from the polyps using D. undulata), Xow speed (9.5 and 21 cm s¡1), and tile with attached a nylon Wlter placed upstream of the tile location. The com- polyps nested within algal treatment. (1b) One-way ANOVA for bination of macroalgal treatment and Xow speed was each Xow speed selected at random from the six possible combinations. (a) ANOVA df MS F-ratio P Each tile with attached polyps was placed in a depression designed to secure each tile to the base of the Xume, and Algal treatment 2 81.205 3.81 0.035 allowed to acclimate in free-stream Xow until ¸3 of the Wve Flow speed 1 17.83 0.837 0.368 X polyps had expanded their tentacles fully. The Wlter was Treatment £ ow speed 2 40.851 1.917 0.167 removed after polyps had expanded. The polyps were Treatment (tile) 27 15.042 0.706 0.815 Wlmed from above during each 5-min experiment with a No Algae versus D. undulata*–– – 0.029 digital video camera (JVC miniDV, Fig. 1). Video images Error 27 21.316 were analyzed to record particle captures using Apple iMo- (b) 9.5 cm/s 1-way ANOVA df MS F-ratio P vie. A capture was deWned as a tentacle retaining a particle for ¸10 s, but in most instances captured particles were Algal treatment 2 10.45 0.698 0.507 retained for the duration of the experiment. In rare Error 27 14.979 – – instances, when macroalgal thalli blocked the video image 21 cm/s 1-way ANOVA df MS F-ratio P of a polyp from above, it was possible to accurately count 0.012 captures by eye from a diVerent angle due to the low num- Algal treatment 2 21.378 5.221 ber of captures (<20 polyp¡1) per experiment, the compara- Error 27 111.607 – – tively large size of Artemia sp. cysts, and the high particle No Algae versus D. undulata*– – – 0.009 retention rate. Particle capture for each treatment was stan- *SigniWcant Tukey post-hoc results 123 Mar Biol (2008) 155:273–280 277

0.1 mm with digital vernier calipers to determine whether polyp size varied among tiles. The mean diameter of each polyp was calculated by averaging the longest and shortest diameters (Chen et al. 1995). Particle Xux past the polyps was calculated for each feeding trial at each Xow speed using the width of the experimental tile (25 mm), the aver- age overall height of polyps and average cyst concentration using the following equation: particle Xux (particles s¡1)= [water velocity (m s¡1) £ cyst concentration (# m¡3)] £ [width (m) £ height of polyp (m)]. Particle Xux is relevant to suspension feeder performance because it represents the number of food particles available for capture if no other structure alters particle paths.

Results Fig. 3 Mean particle capture (§SE) polyp¡1 in each algal treatment at two Xow speeds. Three-one-factor ANOVAs and Tukey post-hoc tests X W V W among algal treatments at each ow speed. Signi cant di erence at Particle capture rate was de ned as the total number of 21 cm s¡1 represented by a and b (adjusted  = 0.017); no signiWcant Artemia sp. cysts captured over a 5-min experimental trial diVerence at 9.5 cm s¡1 divided by the number of expanded C. californica polyps, and each treatment was replicated 10 or 11 times. All parti- Particle Xux past experimental tiles with attached polyps cle capture data were normally distributed and had equal varied between Xow speeds as expected, and ranged variances. The presence of G. robustum, the branching red from 12.0 § 0.4 to 31.0 § 1.4 particles s¡1 (mean § SE, alga, did not signiWcantly aVect particle capture at either Xow Table 2). Particle Xux was higher at 21 cm s¡1 than 9.5 cm speed (Table 1; Fig. 2). In contrast, the presence of the foliose s¡1, accounting for the greater number of particles captured brown alga, D. undulata, signiWcantly inhibited particle by polyps without algae under high versus low Xow capture of C. californica polyps (ANOVA: F2,27 =3.81, trials. The average concentration of Artemia sp. cysts P = 0.035, Table 1; Fig. 2). The interaction between algal (1.6 § 0.1 cysts/ml § SE [n=65]), did not diVer among X W treatment and ow speed was not signi cant (P = 0.167), algal treatments (2-way ANOVA: F2,59 = 0.440, P = 0.646) V X X but particle capture appeared to di er between ow speeds or ow speeds (2-way ANOVA: F1,59 = 2.725, P =0.104), in some algal treatments (Fig. 3). Although the interaction so particle was the same in all three algal treatments within was not signiWcant, posteriori comparisons between Xow each Xow regime (Table 2). speeds were justiWed by a 118% diVerence in particles The retraction behavior of C. californica polyps varied captured at 21 cm s¡1 versus a 19% diVerence in particles signiWcantly among algal treatments (No-Algae [n=11], captured at 9.5 cm s¡1 between the no-algae and D. undulata G. robustum [n=10], D. undulata [n=9]) and between treatments (Fig. 3). To gain further insight into these diVer- Xow speeds (n=30). A greater number of polyps con- ences, a one-way ANOVA with Tukey post-hoc and tracted at 21 cm s¡1 than at 9.5 cm s¡1 (2-way ANOVA:  X adjusted alpha ( = 0.017) was run for each ow speed F1,58 = 6.869, P = 0.011, Fig. 4). The behavior of C. cali- (Table 1). The eVect of D. undulata was signiWcant only at fornica polyps also varied among algal treatments (2-way X W high ow speeds (F2,27 = 5.221, P = 0.012, Tukey post-hoc: ANOVA: F2,58 = 5.096, P = 0.009, Fig. 4). A signi cantly P = 0.009, Table 1). greater number of C. californica polyps contracted in the

Table 2 Particle Xux and particle concentration (§SE) for each algal treatment (No-Algae, G. robustum, D. undulata). Particle density is the aver- age number of cysts per milliliter for each feeding trial Flow speed Particle Xux (cysts s¡1) Particle concentration (cysts ml¡1)

9.5 cm s¡1 21 cm s¡1 9.5 cm s¡1 21 cm s¡1

Algal treatment No-Algae (n = 10) 12.4 § 2.0 33.3 § 3.2 1.5 § 0.25 1.9 § 0.18 G.robustum (n=11) 12.4 § 1.21 29.8 § 3.15 1.5 § 0.15 1.7 § 0.18 D. undulata (n=11) 11.2 § 1.26 28.9 § 2.0 1.4 § 0.16 1.6 § 0.11 All algal treatments (n = 32) 12.0 § 0.39 30.6 § 1.35 1.5 § 0.03 1.7 § 0.09

123 278 Mar Biol (2008) 155:273–280

deXected to entirely cover C. californica tentacles, as occurred frequently for D. undulata thalli. The foliose brown alga, D. undulata, reduced C. californica particle capture by 118% at 21 cm s¡1, and by 64% on average for both Xow speeds. Foliose algae may inhibit particle capture through two distinct mechanisms: (1) redirection of Xow and particles around the polyp tentacles and/or, (2) contact with the polyp that causes tentacle retraction. In a similar study, Dictyota sp., a morphologically similar brown alga to D. undulata, was shown to contact downstream Balano- phyllia sp.*** polyps more frequently than the branching alga, Cystoseira sp. (Coyer et al. 1993). Contact by Dictyota sp. caused the coral to retract its tentacles and allowed Wlamentous and encrusting coralline algae to even- Fig. 4 Percentage of expanded polyps (§SE) during each algal treat- ment and Xow speed. A 2-factor ANOVA and Tukey post-hoc test on tually overgrow the exposed carbonate calyx. In the present arcsin-transformed data among three algal treatments and two Xow study, more C. californica polyps retracted in the presence speeds. SigniWcance among algal treatments represented by a and b of D. undulata than in either the G. robustum or No-Algae (P <0.05) treatments. G. robustum blades consistently remained above the surfaces of C. californica polyps at both Xow D. undulata treatments than in either the G. robustum speeds. Downstream turbulent eddies were observed in (P = 0.018), or No-Algae treatments (P = 0.019, Tukey experimental trials with G. robustum, which could result in post-hoc tests, Fig. 4). a greater number of particles contacting polyps, thereby increasing capture rates (Johnson 2001). Anderson and Charters (1982) demonstrated that at Xow speeds > 6– Discussion 12 cm s¡1,, branches of a morphologically similar species, G. nudifrons, reduced Xow velocities while increasing Our goal in this study was to determine whether the mor- levels of turbulence, which could positively aVect particle phology of macroalgae alters the rate of particle capture by capture success. C. californica, and to elucidate why a signiWcantly higher Capture rates by C. californica within each macroalgal density of C. californica polyps occurs within Eisenia treatment did not diVer signiWcantly between Xow speeds. arborea canopies than outside at Santa Catalina Island, However, graphical analysis indicates that Xow speed may California (Morrow and Carpenter 2008). Using two mor- have altered particle capture rates within algal treatments to phologically distinct macroalgae commonly found in the some degree (Fig. 3). The interaction between algal treat- study habitat, we conclude that diVerent species of macro- ment and Xow speed was not signiWcant most likely algae can mediate particle capture performance and likely because of variability in particle capture among tiles. Over- contribute to the dramatic diVerence in C. californica densi- all, the highest capture rate (mean 12.4 § 1.7 polyp¡1) was ties between canopy habitats. Particle capture was inhibited in the absence of algae at high Xow (21 cm s¡1). Addition- by the foliose brown alga that grows outside of the canopy ally, average particle captures at 9.5 cm s¡1were only 16% (D. undulata) but was not aVected by the branching red lower than at 21 cm s¡1, masking the signiWcant diVerences alga found predominately under the canopy (G. robustum). in capture rates between the No-Algae and D. undulata These results have important implications for the growth treatments at each Xow speed. The Xow environment at and recruitment of many benthic suspension feeders Bird Rock varies from 6 to 70 cm s¡1, suggesting that because they often occur in close association with macroal- polyps of C. californica frequently experience high Xow gae. We determined that a foliose species of macroalgae under natural conditions, and may take advantage of inhibited particle capture performance more than a branch- upstream morphological and topographical features that ing species, therefore, in habitats dominated by similar create turbulent eddies and increase particle capture success taxa, aggregations of suspension feeders may be more com- (R. C. Carpenter unpublished data). In some Xow conditions, mon adjacent to branching algae. associated conspeciWcs or macroalgae acting as topographical The branching red alga G. robustum did not signiWcantly roughness elements could alter particle delivery to suspension alter particle capture by C. californica, and in some cases feeders. the polyps captured more particles when G. robustum was Flow velocity aVects the volume of water passing the present than when no algae were present. This alga may not feeding surface of suspension feeders, the amount of have inhibited particle capture because its thalli rarely particulate material suspended in the water, and the particle 123 Mar Biol (2008) 155:273–280 279 capture eYciency of organisms within the Xow (Koehl is composed primarily of branching red algae (Morrow and 1996). In addition, C. californica polyps contracted more Carpenter 2008). Branching red macroalgae inhibit particle frequently in high Xow than in low Xow. An increase in capture far less than foliose brown algae. Thus, we suggest Xow has been shown to enhance particle capture up to an that the E. arborea canopy sustains an optimal habitat that optimal level, after which it becomes detrimental due to the supports a greater abundance of C. californica polyps. Our deformation of Wltering structures and subsequent tentacle results imply that the morphology of adjacent organisms retraction (Patterson 1984; McFadden 1986; Eckman and can play an essential role in determining which components Duggins 1993; Anthony 1997; Piniak 2002). In the present of the benthic community gain a competitive advantage via study tentacle retraction was higher in the D. undulata enhancement or inhibition of particle capture. treatments compared to the G. robustum and No-Algae treatments, suggesting that D. undulata has negative eVects Acknowledgments This study was supported with funding by the on particle capture through multiple mechanisms. The D. PADI foundation, PADI AWARE, and the California State University, Northridge Graduate Student Association. I am grateful to K. Benes for undulata thalli may inhibit particle capture by contacting dive support and Gerry Smith for technical support at the University of the polyp surface, causing tentacle retraction, and by redi- Southern California (USC) Wrigley Institute for Environmental Stud- recting particles away from polyp tentacles, eVectively ies on Santa Catalina Island. Thank you to Myron and Rich at the Cal- decreasing encounter rates. ifornia State University, Northridge Science Shop for maintenance and construction of necessary equipment. Discussions with N. Chadwick, Benthic invertebrates often occur in environments domi- S. Dudgeon, P. Edmunds, M. Steele, R. Ritson-Williams, and P. Wil- nated by wave-driven oscillatory Xow. As such, future son were essential to the development of this study. The friendship of research would beneWt from studies on the eVects of bi- K. Benes, C. Chabot, R. Elahi, A. Hettinger, B. Kordas, and D. Wang directional Xow on particle capture in the presence of mac- made this project more enjoyable. The experiments described in this study comply with the laws of the United States of America. This is roalgae. Fluid re-sampling could alter capture rates since contribution number 146 of the Marine Biology Program of California oscillatory Xow passes the same particles back and forth State University, Northridge. across the feeding surface (Hunter 1989). The C. califor- nica polyps in this study did not extend >3-cm oV the sub- stratum, and probably would not experience reorientation References of tentacles causing self-shading within aggregations due to X oscillating ow. However, foliose macroalgae could be Anderson SM, Charters AC (1982) A Xuid dynamics study of seawater reoriented in oscillatory Xow, thereby aVecting particle Xow through Gelidium nudifrons. Limnol Oceanogr 27:399–412 redistribution and tentacle retraction. Further studies also Anthony KRN (1997) Prey capture by the Metridium V X would beneWt from experiments at higher Xow speeds (e.g., senile (L.): e ects of body size, ow regime, and upstream ¡1 neighbors. Biol Bull 192:73–86. doi:10.2307/1542577 >30 cm s ) that include variation in the distance between Chadwick NE (1987) InterspeciWc aggressive behavior of the coral- the experimental algal thalli and polyps. A larger distance limorpharian Corynactis californica (: Anthozoa): eVects might allow upstream algal thalli to cause Xow dampening of sympatric corals and sea anemones. Biol Bull 171:110–125. and the creation of downstream turbulent eddies, enhancing doi:10.2307/1541866 Chadwick NE, Adams C (1991) Locomotion, asexual reproduction, particle capture and reducing polyp retraction. and killing of corals by the corallimorpharian Corynactis califor- Suspension feeding is ubiquitous in marine systems, and nica. Hydrobiology 216/217:263–269. doi:10.1007/BF00026473 many studies have addressed the eVects of water Xow on Charters AC, Neushul M, Barilotti C (1968) The functional morphol- particle capture (Rubenstein and Koehl 1977; Hunter 1989; ogy of . In: Margalef R (ed) Proceedings of 6th International Seaweed Symposium. Santiago de Compostela, Patterson 1991; Sebens and Johnson 1991; Helmuth and Spain, pp 89–105 Sebens 1993; Anthony 1997; Sebens et al. 1997; Sebens Chen CA, Chen CP, Chen IM (1995) Sexual and asexual reproduction et al. 1998), but few have addressed the eVects of common of the tropical corallimorpharian Rhodactis (=Discosoma) cohabiting species, such as macroalgae, on particle deliv- indosinensis (Cnidaria: ) in Taiwan. Zool Stud 34:29–40 ery. The present study and our previous experiments Connell SD (2003a) Negative eVects overpower the positive of kelp to (Morrow and Carpenter 2008), demonstrate how sessile exclude invertebrates from the understorey community. Oecologia suspension feeders are impacted by Xow at large and small 137:97–103. doi:10.1007/s00442-003-1312-6 scales within a complex 3-dimensional community. Micro- Connell SD (2003b) The monopolization of understorey habitat by subtidal encrusting coralline algae: a test of the combined eVects habitats created by canopy-forming macrophytes may ben- of canopy-mediated light and sedimentation. Mar Biol (Berl) eWt understory assemblages by mediating Xuctuations and/ 142:1065–1071 or extremes in environmental conditions (e.g., water Xow, Coyer JA, Ambrose RF, Engle JM, Carroll JC (1993) Interactions be- light, sedimentation) (Koehl 1977; Fonseca et al. 1982; tween corals and algae on a temperate zone rocky reef: Mediation by sea urchins. J Exp Mar Biol Ecol 167:21–37. doi:10.1016/ Koehl & Alberte 1988; Koehl 1996). Through the media- 0022-0981(93)90181-M tion of light and Xow, an Eisenia arborea canopy creates an Eckman JE, Duggins DO (1991) Life and death beneath macrophyte understory that is less densely populated by macroalgae and canopies: eVects of understory kelps on growth rates and survival 123 280 Mar Biol (2008) 155:273–280

of marine benthic suspension feeders. Oecologia 87:473–487. Xat blade morphologies. Mar Biol (Berl) 99(3):435–444. doi:10.1007/BF00320409 doi:10.1007/BF02112137 Eckman JE, Duggins DO (1993) EVects of Xow speed on growth of Koehl MAR (1996) When does morphology matter? Annu Rev Ecol benthic suspension feeders. Biol Bull 185:28–41. doi:10.2307/ Syst 27:501–542. doi:10.1146/annurev.ecolsys.27.1.501 1542128 McFadden CS (1986) Colony Wssion increases particle capture rates of Eckman JE, Duggins DO, Sewell AT (1989) Ecology of understory a soft coral: advantages of being a small colony. J Exp Mar Biol kelp environments I. EVects of kelp on Xow and particle transport Ecol 103:1–20. doi:10.1016/0022-0981(86)90129-2 near the bottom. J Exp Mar Biol Ecol 129:173–187. doi:10.1016/ Morrow KM (2006) Shallow kelp canopies regulate macroalgal and 0022-0981(89)90055-5 invertebrate interactions. M.S. Thesis, California State University Fonseca MS, Fisher JS, Zieman JC, Thayer GW (1982) InXuences of Northridge, Northridge the seagrass Zostera marina L on current Xow. Estuar Coast Shelf Morrow KM, Carpenter RC (2008) Shallow kelp canopies mediate Sci 15:351–354. doi:10.1016/0272-7714(82)90046-4 macroalgal composition. EVects on the distribution and abun- Graham LE, Wilcox LW (eds) (2000) Algae. Prentice Hall, Upper dance of Corynactis californica. Mar Ecol Prog Ser, Corallimor- Saddle River pharia. doi: 10.3354/meps07371 Helmuth ST, Sebens KP (1993) The inXuence of colony morphology Okamura B, Partridge JC (1999) Suspension feeding adaptations to and orientation to Xow on particle capture by the scleractinian extreme Xow environments in a marine bryozoan. Biol Bull coral Agaricia agaricites (Linnaeus). J Exp Mar Biol Ecol 196:205–215. doi:10.2307/1542566 165:251–278. doi:10.1016/0022-0981(93)90109-2 Patterson M (1984) Patterns of whole colony prey capture in the octo- Holts LJ, Beauchamp KA (1993) Sexual reproduction in the coral- coral, Alcyonium siderium. Biol Bull 167:613–629. doi:10.2307/ limorpharian sea anemone Corynactis californica in a central 1541414 California kelp forest. Mar Biol (Berl) 116:129–136. doi:10.1007/ Patterson M (1991) The eVects of Xow on polyp-level prey capture in BF00350740 an octocoral, Alcyonium siderium. Biol Bull 180:93–102. Hunter T (1989) Suspension feeding in oscillatory Xow: The eVect of doi:10.2307/1542432 colony morphology and Xow regime on plankton capture by the Piniak GA (2002) EVects of symbiotic status, Xow speed, and prey type hydroid Obelia longissima. Biol Bull 176:41–49. doi:10.2307/ on prey capture by the facultatively symbiotic temperate coral 1541887 Oculina arbuscula. Mar Biol (Berl) 141:449–455. doi:10.1007/ Hurd CL, Stevens CL (1997) Flow simulation around single- and mul- s00227-002-0825-6 tiple-bladed seaweeds with various morphologies. J Phycol Rubenstein DI, Koehl MAR (1977) The mechanisms of Wlter feeding: 33:360–367. doi:10.1111/j.0022-3646.1997.00360.x some theoretical considerations. Am Nat 111:981–994. doi:10.1086/ Irving AD, Connell SD (2006) Predicting understorey structure from 283227 the presence and composition of canopies: An assembly rule for Sebens KP (1981) The allometry of feeding, energetics, and body size marine algae. Oecologia 148:491–502. doi:10.1007/s00442-006- in three sea anemone species. Biol Bull 161:152–171. doi:10.2307/ 0389-0 1541115 Jackson JBC (1977) Competition on marine hard substrata: the adap- Sebens K, Johnson AS (1991) EVects of water movement on prey cap- tive signiWcance of solitary and colonial strategies. Am Nat ture and distribution of reef corals. Hydrobiology 226:91–101. 111:743–767. doi:10.1086/283203 doi:10.1007/BF00006810 Johnson AS (2001) Drag, drafting, and mechanical interactions in can- Sebens KP, Witting J, Helmuth B (1997) EVects of water Xow and opies of the red alga, Chondrus crispus. Biol Bull 201:126–135. branch spacing on particle capture by the reef coral Madracis mir- doi:10.2307/1543328 abilis (Duchassing and Michelotti). J Exp Mar Biol Ecol 211:1– Jompa J, McCook (2002) EVects of competition and herbivory on 28. doi:10.1016/S0022-0981(96)02636-6 interactions between a hard coral and a brown alga. J Exp Mar Sebens KP, Grace SP, Helmuth B, Maney EJ Jr, Miles JS (1998) Water Biol Ecol 271:25–39. doi: 10.1016/S0022-0981(02)00040-0 Xow and prey capture by three scleractinian corals, Madracis mir- Kennelly SJ (1989) EVects of kelp canopies on understorey species due abilis, Montastea cavernosa, and Porites porites in a Weld enclo- to shade and scour. Mar Ecol Prog Ser 50:215–224. doi:10.3354/ sure. Mar Biol (Berl) 131:347–360. doi:10.1007/s002270050328 meps050215 Trager G, Achituv Y, Genin A (1994) EVects of prey escape ability, Knowlton N, Jackson JBC (2001) The ecology of coral reefs. In: Bert- Xow speed, and predator feeding mode on zooplankton capture ness MD, Gaines SD, Hay ME (eds) Marine community ecology. by barnacles. Mar Biol (Berl) 120(2):251–259. doi:10.1007/ Sinauer Associates Inc, Sunderland, pp 395–422 BF00349685 Koehl MAR (1977) EVects of sea anemones on the Xow forces they Vogel S, LaBarbera M (1978) Simple Xow tanks for research and encounter. J Exp Biol 69:87–105 teaching. Bioscience 23:638–643. doi:10.2307/1307394 Koehl MAR, Alberte RS (1988) Flow, Xapping, and photosynthesis of Nereocystis leutkeana: a functional comparison of undulate and

123