effects of chiton granosus (frembly, 1827) and other molluscan grazers on algal succession in wave...

y and Ecology 349 (2007) 84–98www.elsevier.com/locate/jembe

Journal of Experimental Marine Biolog

Effects of Chiton granosus (Frembly, 1827) and other molluscangrazers on algal succession in wave exposed mid-intertidal

rocky shores of central Chile

Moisés A. Aguilera, Sergio A. Navarrete ⁎

Estación Costera de Investigaciones Marinas, Las Cruces, and Center for Advanced Studies in Ecology and Biodiversity,Pontificia Universidad Católica de Chile, Casilla 114-D, Santiago, Chile

Received 15 January 2007; received in revised form 28 April 2007; accepted 7 May 2007

Abstract

Molluscan grazers can have important effects on the abundance, colonization rates, and successional pathways of algalassemblages and the entire intertidal community. In general, early successional algae are more readily consumed than corticated algaeand kelps, which usually get established later in the community succession. To generalize, however, the effect of different grazers onalgal assemblages must be examined on different coasts and under different scenarios. This information could help us understand themechanisms of ecosystem processes and situations in which general models do not apply. Along the coast of Chile, humans harvestlarge keyhole limpets, which seem to be the only invertebrate grazers capable of controlling the dominant corticated algaMazzaellalaminarioides, a canopy-forming species that can cover extensive areas of the mid intertidal zone. In this scenario, where large limpetsare harvested, the overall effects of the diverse molluscan assemblage of limpets, chitons and snails on algal succession and oncorticated algae in particular are not clear. We conducted a 26-month-long experiment to evaluate the effects of molluscan grazers onmid-intertidal algal succession and to isolate the effects of Chiton granosus, the most conspicuous member of the assemblage at thesetidal elevations. At sites heavily impacted by humans themolluscan grazer assemblage had strong negative effects on colonization andabundance of green algae such as ulvoids and Blidingia minima. In doing so, the grazer assemblage had a strong negative indirecteffect on the establishments of chironomid fly larvae, which were only observed on green algal mats and rarely on bare rock. Nosignificant effects were detected on epilithic microalgae, and effects on sessile invertebrates were highly variable over space and time.C. granosus also had significant negative effects on green algae but did not account for the total grazing pressure exerted by the guild.Limited foraging excursions (ca. 35 cm) from refuges and moderate site (crevice) fidelity in this species may contribute to thepatchiness in green algal distribution observed in the field. Nearly 13 months after rock surface were experimentally cleared,M. laminarioides appeared in all experimental plots, but increased over three times faster in enclosures containingC. granosus than inexclosures plots or controls, suggesting that moderate levels of herbivory could actually facilitate the establishment of this alga in thesuccession and that the green algal cover found in the absence of grazers may delay its establishment.© 2007 Elsevier B.V. All rights reserved.

Keywords: Ecological redundancy; Epilithic algae; Foraging behaviour; Grazing; Refuges; Succession

⁎ Corresponding author. Tel.: +56 35 431670; fax:+56 35 431720.E-mail address: [email protected] (S.A. Navarrete).

0022-0981/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jembe.2007.05.002

1. Introduction

Experimental manipulations in different parts of theworld have demonstrated that invertebrate herbivores

mailto:[email protected]

http://dx.doi.org/10.1016/j.jembe.2007.05.002

85M.A. Aguilera, S.A. Navarrete / Journal of Experimental Marine Biology and Ecology 349 (2007) 84–98

can determine the structure and diversity of intertidalalgal assemblages by controlling the abundance (bio-mass, cover) of established dominant macrophytes, bypreventing or severely decreasing colonization by algaeand sessile invertebrates, or by selectively removingedible algal species and, through this, altering succes-sional pathways (see Lubchenco and Gaines, 1981;Hawkins and Hartnoll, 1983; Santelices, 1990; Branchand Moreno, 1994; Fernández et al., 2000; Duffy andHay, 2001 for reviews). Along most temperate coasts ofthe world, the intertidal grazer assemblage is numericallydominated by several species of mollusks and crusta-ceans (Lubchenco and Gaines, 1981; Brawley, 1992;Duffy and Hay, 2001). Several studies have shown thatmost of these grazers prefer early successional algae overthe more herbivore-resistant corticated algae and kelps(sensu Steneck and Dethier, 1994). These algal groupsusually become established later in succession. Excep-tions to these broad generalizations abound, however.Algal responses to grazers depend, among other things,on a) attributes of the grazer assemblage, such as speciescomposition, overall size and the range of feedingpossibilities, which is in part determined by the radularapparatus (Steneck and Watling, 1982), b) attributes ofthe algal species composition and their chemical ormechanical defenses (e.g. Branch, 1981; Gaines, 1985;Hay and Fenical, 1988), c) the rates of algal growth,which are usually determined by external environmentalfactors (Underwood and Jernakoff, 1981; Nielsen andNavarrete, 2004; Coleman et al., 2006; Wieters, 2005),and d) the relative effects of stress gradients on algalversus grazer species (Menge and Olson, 1990; Jenkinsand Hartnoll, 2001; Menge et al., 2002). Since humansare dramatically reducing the biomass of many grazersand top predator species, targeting mostly the largestspecies in these assemblages (Botsford et al., 1997;Jackson et al., 2001; Worm et al., 2002), there is anurgent need to understand how biodiversity and speciescomposition determines the dynamics of benthic com-munities (Coleman et al., 2006). Here, we evaluate theeffect of an entire molluscan grazer assemblage on algalcolonization and succession in an intertidal ecosystemstrongly influenced by humans on the central coast ofChile.

Several intertidal mollusks can control the abun-dance of corticated algae and kelps (Lubchenco, 1978;Moreno and Jaramillo, 1983; Hawkins et al., 1992;Bustamante et al., 1995), but many others can feedefficiently only on benthic microalgae. While they havethe capacity to diminish considerably or even totallyeliminate the thin algal films (diatoms, sporelings andcyanobacteria), they have virtually no direct effects on

established macroalgae (Castenholz, 1961; Nicotri,1977; Underwood, 1984; Hill and Hawkins, 1991;Hawkins et al., 1992). It has also been observed thatsome molluscan grazers can affect the settlement andrecolonization of sessile invertebrates (mostly barna-cles), by eating or exerting a bulldozing effect on larvaeor post-metamorphic stages (e.g. Lottia pelta and Tec-tura scutum Dayton, 1971; Berlow and Navarrete,1997; Scurria (=Collisella) digitalis Paine, 1981; Pa-tella vulgata Hill and Hawkins, 1991; Katharinatunicata Wootton et al., 1996; Collisella grata Chanand Williams, 2003). These differences are primarilydetermined by mechanical restrictions of the feedingapparatus and individual body size. Moreover, differentkinds of molluscan grazers, even those with similarfeeding apparatus and sizes can influence spatialpatterns of algal and sessile invertebrate abundance indifferent ways because of differences in foragingbehavior (see Creese and Underwood, 1982; Chelazziet al., 1987) and the spatial range of foraging excursions(Williams et al., 2000). For example, the mosaics ofsessile species in the intertidal seascape appear to begreatly modified by the distribution of grazers and theirforaging performances, which are directly related tospecific behavioral adaptations (eg. Mackay andUnderwood, 1977; Branch, 1981; Focardi and Chelazzi,1990; Chapman, 2000). In this context, homingbehavior, which is the repeated return of individualsto the same resting site after foraging excursions (seeStimson, 1970; Chelazzi et al., 1987; Focardi andChelazzi, 1990), can have important consequences foralgal community structure and spatial patterns ofvariance (Benedetti-Cecchi, 2000a,b; Coleman et al.,2004).

1.1. The system

A diverse array of herbivores inhabit the mid andupper intertidal zones of the wave-exposed coast ofcentral Chile, including several species of patellid,keyhole and pulmonate limpets, littorinid, turbinid andtrochid snails, as well as several chiton species(Santelices, 1990; Rivadeneira et al., 2002). A fewspecies of grapsid crabs also form part of thisheterogeneous grazing guild, but their abundance isgenerally low in central Chile. At high tide fish also formpart of this guild during some life stages and their effecton macroalgae can be very important, especially in semi-protected areas (Muñoz and Ojeda, 1997). The coast isintensively harvested by humans, who target largeindividuals of the fissurellids, Fissurella crassa andF. limbata (Oliva and Castilla, 1986). Inside the marine

86 M.A. Aguilera, S.A. Navarrete / Journal of Experimental Marine Biology and Ecology 349 (2007) 84–98

reserve of Las Cruces, in central Chile, where noharvesting takes place, Oliva and Castilla (1986)suggest that large keyhole limpets are responsible forthe low abundance of the corticated alga Mazzaellalaminarioides as well as green algae. This result agreeswell with results inside a marine reserve in southernChile involving another large keyhole limpet species(Jara and Moreno, 1984). On the other hand, at sites incentral Chile impacted by human harvesting, andtherefore with low abundances of large limpets, Nielsenand Navarrete (2004) observed that molluscan grazerscan exert strong control on the abundance of ephemeral(mostly green) algae that settled soon after the rocksurface is cleared. However, they also noted that thesegrazers had no significant effects on late successionalspecies, such as Mazzaella, regardless of variation innutrient loadings produced by coastal upwelling.Similarly, Otaíza (1986) found only transient effects onephemeral algae exerted by Chiton granosus, andvirtually no effects on corticated algae. All of thesestudies were conducted on areas of relatively smoothrock surface, away from crevices and therefore poten-tially beyond the reach of most grazers as most non-fissurellid grazers, especially chitons, tend to aggregatein crevices and forage some distance around them.Therefore, their effects on benthic algae might have beenunderestimated in these experiments. Moreover, obser-vations in southern Chile suggest that some of the samegrazer species can control the abundance of corticatedalgae (Jara and Moreno, 1984; Moreno and Jaramillo,1983), in contrast to Nielsen and Navarrete's (2004)results.

Here we explore whether small molluscan grazersplay a role in regulating algal and barnacle colonizationin areas of the coast heavily exploited by humans.We concentrate on the potential effects of the chitonC. granosus, the most abundant grazer in the mid inter-tidal zone in terms of biomass (Otaíza and Santelices,1985), which occurs in crevices during low tides.

2. Methods

2.1. Study site and grazers assemblage

The study was conducted in the mid intertidal zone ofa wave-exposed site outside the Estación Costera deInvestigaciones Marinas (ECIM) of the PontificiaUniversidad Católica de Chile in the locality of LasCruces (33°30′S, 71° 30′W), central Chile. Las Cruces islocated between two upwelling centers and is itselfwithin an “upwelling shadow” (Wieters et al., 2003;Narvaéz et al., 2004; Nielsen and Navarrete, 2004),

where sea surface temperatures remain higher andnutrients lower than at upwelling centers to the northand south (Nielsen and Navarrete, 2004; Wieters, 2005).The experimental site is characterized by exposedbenches oriented SE–NW and separated by widechannels with strong water flux.

To quantify invertebrate herbivore abundance at theexperimental site, we conducted 3 transects perpendicularto the coastline from the highest intertidal fringe (littorinidzone), to lowest intertidal fringe (Lessonia–Corallinezone), spaced ca. 50 cm apart. Along each transect wepositioned 9 to 12 625 cm2 (25×25 cm) quadrats at 30 cmintervals. Sampling was conducted three times during thefirst year of the study (April, September and November2003). Biomass of each species was estimated throughwet weight to length regressions.

2.2. C. granosus diet

To characterize the diet of C. granosus through theyear and relate it to experimental results, we examinedstomach contents of individuals collected monthly tobimonthly in the field, several meters away from theexperimental areas. Preliminary samples and rarefactioncurves (Gotelli and Colwell, 2001) indicated that asample size of 8 individuals was sufficient to charac-terize diet diversity of this species on each samplingdate. Individuals were collected manually from the midand high intertidal zones during night time low tides,injected immediately with a 10% formaldehyde solutionto stop the digestion process, labeled, and taken to thelaboratory for analysis. The gut contents of eachindividual were examined under a dissecting scope toidentify food items to the lowest possible taxonomiclevel. Furthermore, to estimate the relative abundance ofalgal items, we placed a 1 ml gut sample on a glass slidereticulated with 30 points, recording the number ofpoints intersected by each item. Note that more than onefood item could intersect a given point (e.g. microalgaeand sections of macroalgae), and therefore the totalabundance can be higher than 1.

2.3. Movement and foraging activity of C. granosus

To assess the spatial scale of the foraging excursionsof C. granosus from the aggregation crevices andquantify potential homing behavior, we monitored themovement of marked individuals in November 2002(n=50 individuals) and again in February 2003 (n=67individuals) at the same site where the herbivoryexperiments where later set up (see below) and atanother site 500 m away, respectively. The spatial


positions of individuals found inside crevices during theday were recorded on an X–Y coordinate system andindividuals were marked in situ with bee tags withoutprying them off the substratum. A total of 40 markedindividuals were lost in the first site and 34 in the secondone (due to lost tags, migration of the survey area andsome to predation in successive days), leaving 10 and 33individuals respectively at each site. High tag losswas the tradeoff of marking individuals in situ to causethem as little disturbance as possible. To set up thecoordinate reference, we fixed two screws parallel to theshoreline and 5 m apart, along which we extended ameasuring tape to register the position of individualsalong the X-axis. With another measuring tape, extendedperpendicular to the first axis, we recorded the positionof the individuals along the Y-axis. Individuals werefollowed every 12 h, during diurnal and nocturnal lowtides for seven consecutive days. Movement wascalculated as the minimal Euclidian distance (diagonal)between consecutive positions on the X–Y coordinates.The “activity phase” was defined as the change inposition between the diurnal initial position (Dj), thenocturnal position (Nj), and the following diurnalposition (Dj+1) (see Focardi and Chelazzi, 1990). Toquantify the degree of homing, for each activityphase and for each individual, we computed the ratiot=2d/(lo+ lr) (Chelazzi et al., 1987; Focardi andChelazzi, 1990), where lo is the length of the outwardexcursion from the original place, lr is the length ofreturn branch of feeding excursion, and d is theEuclidian distance between Dj and Dj+1.

2.4. Herbivory experiment

In preliminary experiments we evaluated the perfor-mance of different exclosure–enclosure methods formolluscan grazers (a combination of copper-basedantifouling paint, epoxy putty and copper plates) andthe effect of transplanting individuals into experimentalplots to determine if the manipulation altered thepermanence and movement of chitons (see Chapman,2000). We determined that when large keyhole limpetsare naturally absent, the best barrier was provided by a5 cm band of epoxy-putty painted regularly withantifouling paint, applied around natural aggregationsof chitons found inside crevices. Transplanting indivi-duals into smooth rock plots significantly increasedmovements and emigrations rates (see discussions inUnderwood, 1989; Chapman, 1986, 2000). Wire fences(e.g. Coleman et al., 2006) were not used because theycould have interfered with fish grazing and we wanted tofocus on the effects of mollusks.

In line with these preliminary results, we selected25⁎25 cm plots containing crevices with at least 12 to15 adult C. granosus individuals (≥10 mm. at 5th platewidths, see Santelices et al., 1986). Plots were thencleaned of sessile and mobile organism by scraping,brushing and then burning the rock with a propanetorch. The chitons found inside the crevices (mainlylarge individuals ≈18 mm at 5th plate width), werecovered with wet cotton and aluminum paper to preventdesiccation of the crevice and damage to the chitons.Once the experimental areas were denuded, we ran-domly assigned the following treatments: (a) enclosureof C. granosus; in which 12–15 individuals were leftinside each plot and those in excess were removed fromthe plots. Other grazers were also removed (limpets andlittorinids). (b) Exclosure; in which chitons and all othergrazers were removed from the plots. (c) Control: inwhich all grazers were left in position and the plots weremarked only in the corners with epoxy-putty. To controlfor the potential artifact introduced by the “exclosuremethod” we included a (d) Procedural control; in whichgrazers were undisturbed, but the corners of the plotwere delimited with a discontinuous barrier of epoxyputty and antifouling paint, allowing free access tomollusks (see Nielsen and Navarrete, 2004). The surfaceslope and crevice depth and width were recorded foreach experimental plot.

The high heterogeneity of the study site, and theconstraints of finding plots with crevices and naturalaggregation of chitons, made it impossible to place allreplicates within a single homogeneous platform. Instead,the plots were installed in three platforms with slightlydifferent wave exposures and separated between 10–50 m. We used a replicated block design (Kuehl, 1992),blocking by platform and deploying two replicates of eachtreatment within each of the three blocks.

Monthly from November 2002 to April 2003 androughly every two months from then to December 2004,we recorded the cover of sessile organism using a25⁎25 cm quadrat with 81 uniformly spaced intersec-tion points. The numbers of chitons and other grazersfound inside the experimental plot were counted. Whennecessary, individuals were removed to maintain thetreatments. This was done every 5–15 days during thefirst 6 months of the experiment and monthly thereafteras reinvasion of the plots was proven to be low (seeResults section ). To detect micro-spatial (cm) variationsin the colonization of algae and sessile invertebrates inrelationship to crevices, we recorded the cover of sessileorganism using three 5⁎5 cm quadrats (subsamples)placed in a zone “adjacent to crevice” (within 5 cm), andthree in a zone “far from the crevice” (10–15 cm away).

Fig. 1. Mean biomass (g/m2 ±SD) of the most abundant molluscangrazers in the study site at Las Cruces.


2.5. Microalgal abundance

Three times between May and June 2003 weevaluated the effect of C. granosus and other grazerson microalgal (epilithic) biomass by deploying one6.5⁎4.5⁎0.5 cm thick acrylic plate fixed with astainless-steel screw. To facilitate limpet access to platesurface, the sides of these plates were filed in 45° angle,leaving a vertical edge of 0.2 cm. Plates were affixed tothe rock with a flat-head screw flushed with the platesurface and placed in one randomly chosen replicate plotof the enclosure, exclosure and control treatments withineach block. The design was therefore a standardcomplete block design (Kuehl, 1992). Artificial sub-strates have been effectively used before to quantifyepilithic biomass (see Bustamante et al., 1995). Our ownpreliminary observations showed that an acrylic surfaceroughened with sandpaper (300 μm) generates a suitableand homogeneous surface for microalgae. Moreover, thethin profile of the plate allowed herbivores to graze freelyon the plate surface (authors' personal observations).Every 15 days we removed the plates rinsed them with1 ml filtered water, wrapped them in labeled aluminumfoil to prevent contact with light, and deployed cleanplates in the plots. Plates where taken to ECIM labo-ratory to be analyzed immediately or frozen in −20 °Cfor maximum 5 days before analysis. Chlorophyl-aextraction was done following the method proposed byThompson et al. (1999). Briefly, plates were placed incomplete darkness with 50 ml of 100% methanol. After10–15 h they were removed and 1 ml sample was takenand diluted in 5 ml of 100% methanol before recordingtotal fluorescence in a calibrated fluorometer (TurnerDesign AU-10) at 665–750-nm. Data were transformedto μg of chlorophyll-a/cm2 (HMSO, 1986).

2.6. Statistical analyses

Cover of the main sessile species and of bare rockobserved in the experimental plots were analyzed usingseparate univariate repeated measures ANOVA, using areplicated blocked ANOVA as the between-subjectsdesign (Kuehl, 1992). Blocks were considered randomand treatment as fixed. In all cases raw data wereanalyzed because they best met assumptions of normal-ity and variance homogeneity. The Greenhouse–Geissercorrection was used to adjust probabilities when data didnot meet the sphericity assumption for univariate tests ofrepeated measures (Crowder and Hand, 1990; Von Ende,1993). When significant differences among treatmentlevels were found (between subjects), we conducted thefollowing orthogonal planned contrasts: 1) to evaluate

artifacts of manipulation we compared controls versusprocedural controls, 2) to evaluate the effect of all grazers(total herbivory) we compared the control and proce-dural control versus exclosure, 3) to evaluate the effectsof chitons in absence of other grazers we compare theenclosure versus exclosures.

The 5×5 cm subsamples of cover taken at twodistances from the crevices were analyzed with a split-plot complete block ANOVA design with replication inmain plot, and considering the factor distance to creviceas the sub-plot treatment (Kuehl, 1992). To simplify theanalysis we used the average of the three measures foreach distance within each experimental plot. Results oftotal extracted Chlorophyll-a from the experimentalplots used to quantify effects on epilithic algae wereaveraged over the three repetitions, and then analyzedusing a standard randomized complete blocks ANOVA.

The slope of the experimental plot surface, crevicedepth and width were not significantly correlated withany of the response variables examined and thus, theywere not used as covariates in statistical analyses.

3. Results

Field sampling at the study site showed that the mostabundant grazers in terms of numbers and biomass in themid-intertidal zone were the chiton C. granosus andsecondarily the pulmonate limpet Siphonaria lessoniand the scurrinid limpet Scurria zebrina (Fig. 1). Thelittorinid snailEchinolittorina peruviana is also commonbut reach higher biomass in the higher intertidal zone(Fig. 1). Juveniles of the keyhole limpet F. crassa andF. limbata were present at low densities in the mid and

Fig. 2. a) Average relative abundance (±SE) and b) percentage ofindividuals (n=8 per month) found eating the most common fooditems, based on stomach contents of Chiton granosus through the year.

Fig. 3. a) Frequency distribution of the average daily displacements(n=7 days) of 33 individuals of Chiton granosus, and b) frequencydistribution of t parameter which shows the degree of homing throughdiscrete recording of individual position. t value≈0.0 shows perfecthoming, t values≈2.0 show dispersive patterns (Focardi and Chelazzi,1990).


low zones, but adults were rarely observed, apparentlybecause of harvesting by humans (Oliva and Castilla,1986). The most notable non-molluscan macro-grazer inthis habitat were the clingfish Sicyases sanguineus andthe blennid Schartictys viridis. Clingfish are commonlyobserved in vertical walls, but rarely venture onto flatplatforms (Cancino and Castilla, 1988; Loot et al., 2005).The territorial blennid fish is an important grazer in semi-protected channels and pools (Muñoz and Ojeda, 1997),but it is unclear the extent to which this species venturesonto wave exposed benches.

3.1. C. granosus diet

A wide variety of macroalgae, mainly germinativephases and spores and benthic microalgae were foundin gut contents of C. granosus. Macroalgal sporelingswere mostly non-calcareous crusts (Hildebrandialecanelleri and Ralfsia spp), ulvoids (Ulva spp.)and Blidingia minima. Sporelings of Centrocerasclavulatum and Polysiphonia sp were also observed.Moreover, we recorded 11 genera and 13 different spe-cies of diatoms, and two genera of blue green algae(Gomphosphaeria and Plectonema). Fleshy crusts,diatoms and, to a lesser extent ulvoids were the mostabundant items in stomach contents (Fig. 2a). The most

frequent food items were diatoms and calcareous crusts,which were present in the diet throughout the year(Fig. 2b). Cyprids and post-metamorphic stages b3 mmdiameter of barnacles Jehlius cirratus, Notochthamalusscabrosus and Notobalanus flosculus were alsorecorded in the gut content a few months of the year,reaching 100% frequency in October and November2002 (Fig. 2b). Up to 19.4 cyprids (±5.0)/individualand 12.6 (±15.0) post-metamorphic/individuals wereobserved in October and November 2002, respectively.

3.2. Movements and foraging activity of C. granosus

Daily foraging excursions of 33 marked C. granosuswere recorded for seven consecutive days. Other markedchitons were considered inactive or lost during thestudy. Average foraging excursions of 34.3 cm/12 h(SE=32.43) were registered, with a median of 33.3 cm.Approximately 85% of the individuals showed dis-placements of less than 55 cm and only 3.3% of theindividuals moved more than 1 m (Fig. 3a). We did notfind a significant relationship between individual body


size and displacements(R2 =0.004, p=0.7390). More-over, 55.5% of the homing “t” statistic (see Focardi andChelazzi, 1990), was between 0 and 0.8 (where zero isperfect homing, Fig. 3b). Of all individuals considered35% of them had a random pattern of movements withrespect to origin (t value near 1.4, see Focardi andChelazzi, 1990). These results suggest that C. granosushas a moderate home fidelity with a tendency to comeback to resting refuges (crevices) after foraging excur-sions of about 30–40 cm twice a day.

3.3. Herbivory experiments: treatment effectiveness

Diurnal and nocturnal surveys of experimental plotsshowed that the method employed was effective atimpeding the emigration of chitons from enclosures, aswell as the immigration of all grazers into exclosureplots (Fig. 4a–c). In controls and procedural controls thepresence of Siphonaria and other molluscan grazers washighly variable and reached a mean abundance of 30individuals per plot during night time observations (480individuals m−2, Fig. 4b, c). Furthermore, a one-wayANOVA on long-term means per plot showed that the

Fig. 4. Number of individuals per experimental plots of a) Chiton granosusscurrinid limpets), recorded through the study in the different treatments.

mean abundance of chitons in control plots was notsignificantly different to the densities maintained inenclosure plots (F=0.29, d.f=1, 2 p=0.6445 Fig. 4a).

3.4. Herbivory experiments: effects on macroalgae andinvertebrates

The exclosure of all grazers allowed the rapid colo-nization by green algae, composed of Ulva compressa,U. intestinalis, U. rigida, (hereafter treated together asulvoids because of the difficulty of separating them in thefield), and B. minima, which reached over 24% and over40% cover, respectively, within four months followingthe exclosure of all grazers (Fig. 5a,b). In contrast, coverof these algae did not exceed 3% in the presence ofgrazers at natural densities (controls and proceduralcontrols) throughout the duration of the experiments. Inthe presence of chitons alone (enclosures), cover ofulvoids and Blidingia reached intermediate levelsbetween controls and exclosures, fluctuating about10% cover throughout the study. Analyses of long-term means of ulvoids and Blidingia covers showedsimilar results (‘between subjects’ in Table 1). That is,

b) Siphonaria lessoni and c) other molluscan grazers (littorinid snails,

Fig. 5. Monthly mean cover (%±SE) per treatment of a) the green algae Ulva intestinalis and U. compressa (pooled as ulvoids), b) the green algaBlidingia minima, c) bare rock, d) the red alga Porphyra columbina e) the corticated red Mazzaella laminarioides and f) the chthamalid barnaclesJehlius cirratus and Nothochthamalus scabrosus (pooled).


significant differences among treatment levels, non-significant differences among blocks and no change intreatment effect across blocks (Table 1). For both algaltaxa, planned contrasts showed that cover in controls andprocedural controls was not different, that they both weresignificantly lower than in grazer exclosures, and thatcover in the presence of chitons alone (enclosures) wassignificantly lower than when all grazers were removed(Table 1 planned contrasts). The time course of algalresponses to grazer treatments differed between ulvoidsand Blidingia. In both cases treatment effects changedsignificantly over time (significant time× treatmentinteraction for within subject effects in Table 1).However, while cover of ulvoids in the grazer exclosuretreatment tended to decrease about 14–15 months after

initiation of the study, cover of Blidingia in theexclosures remained high almost until the end of theexperiment (Fig. 5b). The decrease of Blidingia at theend of the experiment was probably produced byreinvasion of grazers after we reduced the frequency ofgrazer removals, by September 2004.

Treatment effects on the amount of bare rock werehighly variable over time (Table 1). For several months ofthe year, bare rock was lowest in the grazer exclosuretreatment, particularly during the first year of the ex-periment (Fig. 5c), however, during the second year thesedifferences tended to disappear. The effect of all mollus-can grazers led to relatively high levels of bare rock incontrols and procedural controls, but these treatmenteffects varied significantly over time (Table 1, Fig. 5c).

Table 1Results of repeated-measures ANOVAof the herbivory experiment usinga replicated block design, considering blocks as random and treatment asfixed factors (exclosure, inclusion, control and procedural control)

Source df MS F p

A) Bare rockBetween subjectsTreatments 3 12619.249 3.25 0.118Block 2 1704.458 1.04 0.384Treat⁎block 5 3882.121 2.38 0.107Error 11 1634.370

Within subjectsTime 11 2436.827 7.78 b .0001Time⁎ treat 33 519.060 2.24 0.004Time⁎block 22 508.613 1.62 0.051Time⁎ treat⁎block 55 232.203 0.74 0.893Error(time) 121 313.303

B) UlvoidsBetween subjectsTreatments 3 4008.078 18.07 0.004Block 2 185.289 0.89 0.438Treat⁎block 5 221.828 1.06 0.430Error 11 208.528

Othogonal planned Contrasts

Control vs procedural control(PC)

1 6.955 0.41 0.546

Control-PC vs exclosure 1 1008.650 59.23 0.0003Enclosure vs exclosure 1 278.090 16.33 0.007

Within subjectsTime 11 159.755 1.49 0.145Time⁎ treat 33 118.043 1.97 0.012Time⁎block 22 111.366 1.04 0.427Time⁎ treat⁎block 55 59.797 0.56 0.992Error(time) 121 107.530

C) Blidingia minimaBetween subjectsTreatments 3 13390.600 41.81 0.0006Block 2 421.118 0.29 0.751Treat⁎block 5 320.269 0.22 0.945Error 11 1439.014

Othogonal planned Contrasts

Control vs procedural control(PC)

1 0.249 0.01 0.9230

Control-PC vs exclosure 1 3692.354 151.42 b0.0001Enclosure vs exclosure 1 1403.656 57.56 0.0003

Within subjectsTime 11 308.354 4.52 b0.0001Time⁎ treat 33 238.234 3.58 b0.0001Time⁎block 22 139.843 2.05 0.0071Time⁎ treat⁎block 55 66.625 0.98 0.5292Error (time) 121 68.229

Table 1 (continued)

Source df MS F p

D) Mazzaella laminarioides (after 300 days)Between subjectsTreat 3 0.040 1.20 0.399Block 2 0.011 4.59 0.035D) Mazzaella laminarioides (after 300 days)Between subjectsTreat⁎block 5 0.033 13.45 b .0001Error 11 0.002

Within subjectsTime 4 0.018 14.35 b .0001Time⁎ treat 12 0.004 1.10 0.412Time⁎block 8 0.001 0.69 0.694Time⁎ treat⁎block 20 0.004 3.06 0.001Error(time) 44 0.001

E) Chthamalid barnaclesBetween subjectsTreat 3 4708.251 2.02 0.229Block 2 4044.738 5.42 0.023Treat⁎block 5 2326.902 3.12 0.053Error 11 746.138

Within SubjectsTime 11 1788.435 18.21 b .0001Time⁎ treat 33 139.702 1.85 0.021Time⁎block 22 125.606 1.28 0.199Time⁎ treat⁎block 55 75.565 0.77 0.861Error(time) 121 98.233

Separate analyses were performed for cover of A) bare rock,B) ulvoids, C) Blidingia minima, D) Mazzaella laminarioides andE) chthamalid barnacles (Notochthamalus scabrosus, Jehlius cirratus).


There were no treatment effects on other algae, such asPorphyra spp. (Fig. 5d) or fleshy and calcareous crusts.By April 2003, six months after clearing the plots, thepresence of corticated algae, mostly M. laminarioides(see Santelices, 1990), appeared in experimental plots,particularly in chiton enclosures and in the form ofholdfast crusts. However, cover of these formsremained below 1–2% at least until December 2003(Fig. 5e). Only then the upright form of Mazzaellabecame apparent, particularly in chiton enclosures,where it reached over 25% by the end of the experiment(Fig. 5e). In the presence of all grazers, as well as in theabsence of all grazers, Mazzaella cover did not surpass5% by the end of the experiment. Statistical analyses forthis species were therefore restricted to the second partof the experiment, from December 2003 throughDecember 2004. Large variation in rates of increaseamong blocks was reflected in significant treatment× -block interactions for ‘between subjects’ comparisonsand significant time× treatment×block interaction for‘within subjects’ comparisons (Table 1). Estimating the

Fig. 7. Mean chlorophyll-a concentration (ug/cm2±SE) recordedevery 15 days between May and June 2003 (averaged per plot) onacrylic plates (6.5cm⁎4.5 cm⁎0.5cm) installed inside the experimen-tal enclosures, exclosures and controls (see text for details).


rates of increase in Mazzaella (% cover/day) from linearOLS regression for each experimental plot, showed thatrates in the chiton enclosureswere 3.3 times higher than inexclosure plots and 3.9 times faster than in control plots.Analysis of variance on the slopes of these regressions(Mazzella cover/time) for each plot showed significantdifferences among treatments (F=7.14, d.f. =3, 12,p=0.0052), however, these treatment effects on ratesalso vary significantly across blocks (block×treatmentinteraction, F=4.37, d.f.=6, 12, p=0.0143). Takentogether these results suggests that C. granosus canhave a positive direct or indirect effect in the establish-ment of Mazzaella later in succession and that thepresence of ephemeral algae as well as high herbivorypressure exerted by all grazers together might delay theestablishment of this species.

In October 2002, one month after the start of theexperiments, we observed the presence of barnaclespat. The majority of the barnacles were the chthama-lids N. scabrosus and J. cirratus (pooled togetheras Chthamalids), with a few N. flosculus. Cover ofbarnacles in the different treatments was highly var-iable over time and across blocks, rendering significantthe long-term differences among blocks (Table 1‘between subjects’) and the treatment× time interaction(Table 1, ‘within subjects’). Barnacle cover was higherin control and procedural controls, remaining between30 and 40% but switching in relative abundances overtime (Fig. 5f).

In January 2003 (four months since the start of theexperiments) we observed the presence of chironomidfly larvae associated with the green algae in enclosureand exclosure plots (Fig. 6). The maximummean densityof insect larvae was 8.9 (±8.7) larvae/25 cm2 in January2003 in exclosure plots, while they were never found in

Fig. 6. Mean density (individual/625 cm2) of chironomid fly larvaerecorded inside the experimental plots in 5⁎5 cm quadrats (3 replicatesper plot).

controls and only rarely in enclosures (Fig. 6). Nostatistical analyses were performed to compare treat-ments because no larvae were found in controls andprocedural controls. Density of insect larvae was signif-icantly correlated with cover of Blidingia (Spearmanr=0.787, pb0.0001) and ulvoids (r=0.772, pb0.0001).

No significant differences were observed in the coverof Blidingia, ulvoids or bare rock between areasadjacent (first 5 cm) from those far (10–15 cm) fromthe crevice (split-plot Anova, pN0.05 for the treatmentfactor in all cases).

3.5. Herbivory experiments: effects on microalgae

Acrylic plates for microalgal settlement permitted usto evaluate grazer effects on epilithic microalgalbiomass. Field observations showed that all molluscanspecies were able to crawl on the plates (i.e in controlplots), and subsequent observation of acrylic platesunder a dissecting scope evidenced marks of radula onthe surface. However, ANOVA on extracted chloro-phyll-a did not show significant treatment differences(F=0.073, df=2, pN0.05; Fig. 7).

4. Discussion

Our results show that the molluscan grazer assemblagefound at sites heavily impacted by humans has strongnegative effects on colonization and abundance of greenalgae such as ulvoids and B. minima. By doing this, thegrazer assemblage has a strong, negative, indirect effect onthe establishment of chironomid fly larvae, which wereonly observed on green algalmats and rarely found on barerock or other substrates. By itself, the common chiton, C.


granosus, has significant negative effects on green algaebut it does not account for all grazing pressure exerted bythe guild. After nearly 13 months, the corticated alga M.laminarioides appeared in all experimental plots andincreased faster in control plots and especially inenclosures ofC. granosus, suggesting that moderate levelsof herbivory by these grazers could actually favor theestablishment of this alga during later succession. Here wediscuss these results in the light of previous studies and theconsequences for the community.

4.1. Direct and indirect effects of grazing on cover ofgreen algae and barnacles

Our experiments show that small molluscan grazerscan control the colonization and abundance of green algae(ulvoids, Blidingia) in wave-exposed areas. This negativeeffect corresponds well with the diet of C. granosus(see diet results) and that of other grazers (see Santeliceset al., 1986). Indeed, ulvoids and Blidingia were found athigh frequency in stomach contents. Several studies onother coasts have found similar results, i.e. experimentalremoval of molluscan grazers leads to fast growth ofopportunistic, highly palatable algae, which can some-times persist for long time in the absence of the grazers(e.g. Lubchenco, 1978; Lubchenco and Menge, 1978;Underwood and Jernakoff, 1981; Nielsen and Navarrete,2004). In our experiments chitons by themselves, atdensities similar to those observed in the field, hadsignificant negative effects on green algal cover, but didnot account for the near absence of green cover exerted bythe entire molluscan assemblage. The combined action ofchitons, small limpets and snails kept green algal coverbelow 3%, while chitons alone maintained green algalcover at about 10–15%. This result suggests that chitonsand other small grazing mollusks have qualitativelysimilar effects on green algae and that the effects of thesmaller mollusks are not redundant (sensu Lawton, 1994;Allison et al., 1996; Navarrete and Menge, 1996) withthose of chitons. Further studies are needed to determinewhether interaction strengths between grazer species andgreen algae follow a keystone type of distribution (e.g.Paine, 1992; Navarrete and Castilla, 2003; Navarrete andMenge, 1996) orwhether they all contribute about equallyto the total grazing effect. Whether these patterns ofgrazing change in the presence of large fissurellid limpets,for instance inside marine protected areas (Oliva andCastilla, 1986), should also be investigated.

Roughly three months after the beginning of experi-ments we observed chironomid fly larvae on most of theexclosure plots that contained a green cover of ulvoidsand particularly on Blidingia. These larvae are common

and patchily abundant in intertidal algal mats of centraland southern Chile (Camus and Barahona, 2002). Scarceinformation exists about their effects in the intertidalzone to assess their potential role in keeping green algalcover under check. Robles and Cubit (1981) concludedthat dipteran larvae can influence the succession anddetermine the abundance of high intertidal algae incentral California, USA rocky shores (and see Cubit,1984). We observed fly larvae eating ulvoids and Bli-dingia, but judging by their size and low densities, theireffects on algal cover at the study site are probablynegligible. Fly larvae were absent from bare rock andother algal fronds, suggesting that grazers have anegative indirect effect on insect larvae throughcontrolling green algal cover, similar to findings reportedfor the coast of Washington, USA (Wootton et al., 1996).However, we did find insect larvae in the diet of C.granosus, albeit at low frequency (see also Aguilera,2005), and therefore direct negative effects throughconsumption cannot be ruled out.

In experiments conducted in central Chile Otaíza(1986) found moderate effects of chitons on green algalcolonization and no effects on corticated algae, whichwere never observed throughout the year-long experi-ment. Lack of colonization by corticated algae duringOtaiza's experiment agrees well with Nielsen andNavarrete's (2004) and our own results showing thatthese algae become established only about a year afterclearing the rock surface. In his enclosures ofC. granosusOtaíza (1986) found an intensification of grazing effectsin areas of the plots where chitons aggregated. In contrast,we found only weak, non-significant spatial differences ingrazer effects within the enclosures. Since chitons' dailyexcursions from their refuges are about 35 cm, the size ofthe plot might not be large enough to observe spatialvariation of grazing and suggest that chitons can haveeffects on green algae over comparatively large areas.Differences with Otaiza's (1986) results are probablyrelated to the fact that his plots did not include a crevice orrefuge and ours did. Indeed, our observations show thatchitons always forage at high tides and at nights and returnto a refuge (crevice) at low tides and that they exhibitmoderate levels of homing (site fidelity). Therefore theirforaging behavior might be altered when transplanted andplaced on smooth rock.

Many studies have documented that molluscangrazers can control the abundance of barnacles throughreducing post-settlement survival (e.g. Dayton, 1971;Berlow and Navarrete, 1997; Chan andWilliams, 2003).This negative effect can be the result of direct con-sumption or bulldozing of cyprids and post-metamor-phic stages. Our diet analyses of C. granosus showed


that it consumes barnacle cyprids and spat during themonths following barnacle settlement on the shore(October–December), this consumption of barnacle setis also frequently observed in southern and northernChile (Jara and Moreno, 1984; Aguilera, 2005). Highvariability among blocks and over time precluded ageneral conclusion about overall treatment effects ofgrazers on barnacles, however. Considering the relative-ly high barnacle recruitment rates observed at Las Cruces(see Lagos et al., 2005), it is possible that grazers cancontrol barnacle cover only in small areas that receivelower barnacle settlement. The observed differencesmight also be produced by facilitative effects of non-chiton grazers on barnacle settlement, which deservesfurther investigation.

Our results from acrylic plates suggest that grazerscannot control the abundance of epilithic microalgae,which contrasts with findings in other regions (e.g.Hawkins et al., 1992; Bustamante et al., 1995). Thissuggests that rates of renewal of microalgae are too fastto observe negative effects at the densities of molluscsobserved in the study site. However, while we directlyobserved grazers on the acrylic plates in the field andfound radula marks under the microscope, it is possiblethat the rate of grazing on the plates was lower than onthe rock surface.

4.2. Direct and indirect effects on corticated algae

One of the main problems in assessing the effect ofbenthic grazers on corticated algae is that the dominantmid-intertidal species, M. laminarioides, typically onlyappears in experimental plots more than a year afterthe rock surface has been cleared (Otaíza, 1986; Nielsenand Navarrete, 2004). Maintenance of treatments fora sufficiently long time is therefore logistically challeng-ing and makes comparisons of results among differentstudies of varying duration rather confusing. Ourstudy was conducted for over two years, and only after12–13 months were erect Mazzaella fronds observed inexperimental plots. This delayed appearance of corticatedalgae has been observed in other intertidal systems (i.e.Lubchenco, 1978; Lubchenco and Menge, 1978; Bene-detti-Cecchi, 2000a,b). It is entirely possible that otherfactors (e.g. grazing by large fish) restrict corticated algalgrowth (Muñoz and Ojeda, 1997), but at this point wehave no evidence of such effects in direct wave exposedareas. Therefore, the slow appearance of this “latesuccessional” species is likely related to slow initialcolonization of bare rock, which may be aided bycoalescence of spores once initiated (Santelices et al.,2003; Santelices, 2004).

Despite these difficulties, experimental studies con-ducted in southern and central Chile allow someconclusions regarding factors that control Mazzaellaand other mid-intertidal corticated algae. In southernChile Moreno and Jaramillo (1983) found that in theabsence of humans, large Fissurella picta play a keyrole in controllingMazzaella (= Iridaea in that study) inthe mid intertidal rocky shore. Similarly, Oliva andCastilla's (1986) results suggest other large keyholelimpets, F. crassa and F. limbata, can control corticatedalgal cover inside the marine reserve of Las Cruces.Results from areas impacted by humans are lessconcordant. At sites affected by human harvesting insouthern Chile, Jara and Moreno (1984) conclude thatS. lessoni has significant negative effects on Mazzaellaabundance through trimming its fronds. Studies byNielsen and Navarrete (2004) in human-impacted areasof central Chile did not show any effects of grazers onMazzaella, regardless of the level of nutrient inputproduced by coastal upwelling. Our results from human-impacted areas in Las Cruces suggest a rather complexpattern of effects on late successional species that takeonly place more than a year after the clearance of therock surfaces. Colonization and growth of Mazzaellawas slowest in control plots where all grazers werepresent and in exclosure plots with no grazers at all,while the fastest rates were observed in chiton enclo-sures. We interpret this to be an outcome of varyinglevels of grazing intensity and competition with greenalgae. Thus, moderate grazing by C. granosus (assuggested by intermediate cover of green algae) couldfacilitate the colonization of Mazzaella. Foraging bychitons created bare rock (ca. 30%) in the green algalmat that could be colonized later by Mazzaella. Whenall grazers were excluded, both ulvoids and Blidingiapersisted through the study and may have delayed orprevented Mazzaella spores settlement. Studies in othersystems have experimentally shown that grazers canpermit the establishment of corticated algae on emergentsubstrate through elimination of ephemeral algae whichcan competitively inhibit their colonization (e.g. Lub-chenco, 1978; Lubchenco and Menge, 1978; Sousa,1979; Cubit, 1984; Benedetti-Cecchi, 2000a,b). In thepresence of all grazers together, grazing pressure mightbe sufficiently high and spatially uniform to eliminatesporelings of Mazzaella and substantially delay, but notprevent its dominance. Thus, removal or severereduction of chiton abundance, as it could happen ifhumans “fish down the food web” for the smallerbodied herbivores, could lead to an overall reduction inMazzaella. Whether herbivorous fish, which were notexcluded in our experiments, play any role in controlling


abundance of Mazzaella, particularly in the low zone,should also be experimentally investigated (Muñoz andOjeda 1997; Vinueza et al., 2006).

In summary our results show that all non-fissurellidmolluscan grazers play a role in controlling green algalcover in mid-intertidal exposed rocky-shores of centralChile, but they appear to not be able to compensate forthe absence of large fissurellids limpets in controllingthe abundance of late successional Mazzaella. In fact,C. granosus alone can positively affect the establish-ment of Mazzaella during late succession. It is thereforeintriguing what factors prevent Mazzaella from attain-ing higher primary or secondary cover (on musselshells) in the mid and particularly the low shore andfurther studies should concentrate on examining otherbiological and physical factors. Future studies shouldalso concentrate on teasing apart the effects of thedifferent species that compose this diverse assemblageto determine what is the pattern of interaction strengths(e.g. keystone, diffuse) that characterize the guild.

Acknowledgements

We are in debt to a number of friends and students forfield assistance, including Tatiana Manzur, Paolo Tibal-deschi, Mirtala Parragué and Fredy Véliz. Comments bymembers of ECIM lab were very helpful in preparing thismanuscript. This project was funded by grants to SANfrom Fondecyt #1040787 and Fondap-Fondecyt 15001-001, CASEB Center. MAAwas supported by the AndrewMellon Foundation during these studies and by CONICYTscholarship during the writing of the manuscript, for whichwe are most grateful. [RH]

References

Aguilera, M.A., 2005. Cirripedios en la dieta del molusco herbívoroChiton granosus (Frembly, 1827) presente en el intermareal rocosode Iquique, norte de Chile. Invest. Mar., Valparaíso 33, 109–113.

Allison, G.W., Menge, B.A., Lubchenco, J., Navarrete, S.A., 1996.Predictability and uncertainty in community regulation: conse-quences of reduced consumer diversity in coastal rocky ecosystems.In: Mooney, H.A., Cushman, J.H., Medina, E., Sala, O.E., Schulze,E.D. (Eds.), Functional Roles of Biodiversity: a Global Perspective.Wiley Press, Chichester, England, pp. 371–392.

Benedetti-Cecchi, L., 2000a. PredictingDirect and Indirect Interactionsduring Succession in a Mid-Littoral Rocky Shore Assemblage.Ecol. Monogr. 70, 45–72.

Benedetti-Cecchi, L., 2000b. Variance in ecological consumer–resourceinteractions. Nature 407, 370–374.

Berlow, E.L., Navarrete, S.A., 1997. Spatial and temporal variation inrocky intertidal community organization: lessons from repeatingfield experiments. J. Exp. Mar. Biol. Ecol. 214, 195–229.

Botsford, L.W., Castilla, J.C., Peterson, Ch.H., 1997. The managementof fisheries and marine ecosystems. Science 277, 509–515.

Branch, G.M., 1981. The biology of limpets: physical factors; energy flowand ecological interactions. Ocean.Mar. Biol. Ann.Rev. 19, 235–380.

Branch, G.M., Moreno, C.A., 1994. Intertidal and subtidal grazers. In:Siegfield, W.R. (Ed.), Rocky shores: Exploitation in Chile andSouth Africa. Springer, Berlin, pp. 75–100.

Brawley, S.H., 1992. Mesoherbivores. In: John, D.M., Hawkins, S.J.,Price, J.H. (Eds.), Plant-Animal interaction in the Marine Benthos.Systematics Assotiation Special Volume N° 46. Clarendon Press,Oxford, pp. 235–263.

Bustamante, R., Branch, G., Eekhout, S., Robertson, B., Zoutendyk,P., Schleyer, M., Dye, A., Hanekom, N., Keats, D., Jurd, M.,McQuaid, C., 1995. Gradients of intertidal primary productivityaround the coast of South Africa and their relationships withconsumer biomass. Oecologia 102, 189–201.

Camus, P., Barahona, R., 2002. Insectos del Intermareal de Concepción,Chile: perspectivas para la investigación ecológica. Rev. Chil. Hist.Nat. 73, 793–803.

Cancino, J.M., Castilla, J.C., 1988. Emersion behavior and foragingecology of the common Chilean clingfish Sicyases sanguineus(Pisces: Gobiesocidae). J. Nat. His. 22, 249–261.

Castenholz, R.W., 1961. The effect of grazing on marine litoral diatompopulations. Ecology 42, 783–794.

Coleman, R.A., Underwood, A.J., Chapman, M.G., 2004. Absence ofcosts of foraging excursions in relation to limpet aggregation.J. Anim. Ecol. 73, 577–584.

Coleman, R.A., Underwood, A.J., Benedetti-Cecchi, L., Åberg, P.,Arenas, F., Arrontes, J., Castro, J., Hartnoll, R.G., Jenkins, S.R.,Paula, J., Della Santina, P., Hawkins, S.J., 2006. A continental scaleevaluation of the role of limpet grazing on rocky shores. Oecologia147, 556–564.

Cubit, J.D., 1984. Herbivory and the seasonal abundance of algae on ahigh intertidal rocky Shore. Ecology 65, 1904–1917.

Chan, B.K.K., Williams, G.A., 2003. The impact of physical stress andmolluscan grazing on the settlement and recruitment of Tetraclitaspecies (Cirripedia: Balanomorpha) on a tropical shore. J. Exp.Mar. Biol. Ecol. 284, 1–23.

Chapman, M.G., 1986. Assesment of some control in experimentaltransplants of intertidal gastropods. J. Exp. Mar. Biol. Ecol. 103,181–201.

Chapman, M.G., 2000. Poor design of behavioural experiments getspoor results: examples from intertidal habitats. J. Exp. Mar. Biol.Ecol. 250, 77–95.

Chelazzi, G., Della Santina, P., Parpagnoli, D., 1987. Trail following inthe chiton Acanthopleura gemmata: Operational ecologicalproblems. Mar. Biol. 95, 539–545.

Creese, R.G., Underwood, A.J., 1982. Analysis of inter- and intra-specific competition amongst intertidal limpets with differentmethods of feeding. Oecologia 53, 337–346.

Crowder,M.J., Hand,D.J., 1990.Analysis of RepeatedMeasures, vol. 41.CRC publisher.

Dayton, P.K., 1971. Competition, disturbance, and communityorganization: the provision and subsequent utilization of space ina rocky intertidal community. Ecol. Monogr. 41, 351–389.

Duffy, J.E., Hay, M.E., 2001. The ecology and evolution of marineconsumer-prey interactions. In: Bertness, M.D., Hay, M.E.,Gaines, S.D. (Eds.), Marine Community Ecology. SinauerAssociates, Sunderland, MA, pp. 131–157.

Fernández, M., Jaramillo, E., Marquet, P., Moreno, C., Navarrete, S.,Ojeda, P., Valdovinos, C., Vásquez, J., 2000. Diversity, ecologyand biogeography of Chilean benthic nearshore ecosystems: anoverview and needs for conservation. Rev. Chil. Hist. Nat. 73,629–662.


Focardi, S., Chelazzi, G., 1990. Ecological determinants of bioeconomicsin three intertidal of chitons (Acanthopleura spp). J. Anim. Ecol. 59,347–362.

Gaines, S.D., 1985. Herbivory and between-habitat diversity: thedifferential effectiveness of defenses in a marine plant. Ecology 66,473–485.

Gotelli, N.J., Colwell, R.K., 2001. Quantifying biodiversity: proce-dures and pitfalls in the measurement and comparison of speciesrichness. Ecol. Lett. 4, 379–391.

Hawkins, S.J., Hartnoll, R.G., 1983. Grazing of intertidal algae bymarine invertebrates. Ocean. Mar. Biol. Ann. Rev. 21, 195–282.

Hawkins, S.J., Hartnoll, R.G., Kain, J.M., Norton, T.A., 1992. Plant–animal interactions on hard substrata in the north-eastAtlantic. In: John,D.M., Hawkins, S.J., Price, J.H. (Eds.), Plant-Animal interaction in theMarine Benthos. Systematic Association Special volume No. 46.Clarendon Press, Oxford, pp. 1–32.

Hay, M.E., Fenical, W., 1988. Marine plant–herbivore interactions: theecology of chemical defense. Ann. Rev. Ecolog. Syst. 19, 111–145.

Hill, A.S., Hawkins, S.J., 1991. Seasonal and spatial variation ofepilithic microalgal distribution and abundance and its ingestion byPatella vulgata on a moderately exposed rocky shore. J. Mar. Biol.Assoc. UK 71, 403–423.

HMSO, 1986. The determination of chlorophyll-a in aquatic environ-ments.Methods for the examination of water and associatedmaterials,vol. 4, sec 2. HMSO, London, pp. 6–13.

Jackson, J.B.C., et al., 2001. Historical overfishing and the recentcollapse of coastal ecosystems. Science 293, 629–638.

Jara, H.F.,Moreno, C.A., 1984. Herbivory and structure in aMidlittoralRocky Community: a case in southern Chile. Ecology 65, 28–38.

Jenkins, S.R., Hartnoll, R.G., 2001. Food supply, grazing activity andgrowth rate in the limpet Patella vulgata L.: a comparisonbetween exposed and sheltered shores. J. Exp. Mar. Biol. Ecol.258, 123–139.

Kuehl, R.O., 1992. Design of Experiments, statistical principles ofresearch design and analysis, Second edition. Duxbury Press edts,pp. 263–299.

Lagos, N.A., Navarrete, S.A., Véliz, F., Masuero, A., Castilla, J.C.,2005. Meso-scale spatial variation in settlement and recruitment ofintertidal barnacles along central Chile. Mar. Ecol. Prog. Ser. 290,165–178.

Lawton, J.H., 1994. What do species do in ecosystems? Oikos 71,367–374.

Loot, G., Aldana, M., Navarrete, S.A., 2005. Effects of humanexclosure on parasitism in intertidal food webs of central Chile.Conserv. Biol. 19, 203–212.

Lubchenco, J., 1978. Plant species diversity in a marine intertidalcommunity: Importance of herbivore food preferences and algalcompetitive abilities. Am. Nat. 112, 23–39.

Lubchenco, J., Gaines, S.D., 1981. A unified approach to marineplant–herbivore interactions. I. Populations and Communities.Ann. Rev. Ecolog. Syst. 12, 405–437.

Lubchenco, J., Menge, B.A., 1978. Community development andpersistence in a low rocky intertidal zone. Ecol.Monogr. 48, 67–94.

Mackay, D.A., Underwood, A.J., 1977. Experimental studies onhoming in the intertidal patellid limpet Cellana tramoserica(Sowerby). Oecologia 30, 215–237.

Menge, B.A., Olson, A.M., 1990. Role of scale and environmentalfactors in regulation of community structure. TREE 5, 52–57.

Menge, B.A., Olson, A.M., Dahlhoff, E., 2002. Environmental stress,bottom-up effects, and community dynamics: integrating molec-ular-physiological with ecological approaches. Integr. Comp. Biol.42, 892–908.

Moreno, C.A., Jaramillo, E., 1983. The role of grazers in the zonationof intertidal macroalgae of the Chilean coast. Oikos 41, 73–76.

Muñoz, A.A., Ojeda, F.P., 1997. Feeding ecology and guild structureof a rocky intertidal fish assemblage of Central Chile. Environ.Biol. Fish. 49, 471–479.

Narvaéz, D., Poulin, E., Leiva, G., Hernández, E., Castilla, J.C.,Navarrete, S.A., 2004. Seasonal variation of nearshore hydrographicconditions in Central Chile. Cont. Shelf Res. 24, 279–292.

Navarrete, S.A., Castilla, J.C., 2003. Experimental determination ofpredation intensity in an intertidal predator guild: dominant versussubordinate prey. Oikos 100, 251–262.

Navarrete, S.A., Menge, B.A., 1996. Keystone predation andinteraction strength: interactive effects of predators on their mainprey. Ecol. Monogr. 66, 409–429.

Nicotri, M.E., 1977. Grazing effect of four marine intertidal herbivoreson the microflora. Ecology 58, 1020–1032.

Nielsen, K., Navarrete, S.A., 2004. Mesoscale regulation comes fromthe bottom-up: intertidal interactions between consumers andupwelling. Ecol. Lett. 7, 31–36.

Oliva, D., Castilla, J.C., 1986. The effects of human exclosure on thepopulation structure of key-hole limpets Fissurella crassa andFissurella limbata in the coast of Central Chile. P.S.Z.N.I. Mar.Ecol. 7, 201–217.

Otaíza, R.D., 1986. Patrones de distribución vertical de chitones yefecto de pastoreo de Chiton granosus en roqueríos intermarealesChile central. In: Santelices, Bernabé (Ed.), Simposio Internacio-nal. Monografías Biológicas No. 4. Usos y Funciones ecológicasde las algas marinas bentónicas, pp. 73–190.

Otaíza, R.D., Santelices, B., 1985. Vertical distributions of chitons(Mollusca: Polyplacophora) in the rocky intertidal zone of Chile.J. Exp. Mar. Biol. Ecol. 86, 229–240.

Paine, R.T., 1981. Barnacle ecology: is competition important? Theforgotten roles of disturbance and predation. Paleobiology 7, 553–560.

Paine, R.T., 1992. Food-webs analysis through field measurement ofper capita interaction strength. Nature 335, 73–75.

Rivadeneira, M., Fernández, M., Navarrete, S.A., 2002. Latitudinaltrends of species diversity in rocky intertidal herbivore assem-blages: spatial-scale and the relationship between local andregional species richness. Mar. Ecol. Prog. Ser. 245, 123–131.

Robles, C., Cubit, J., 1981. Influence of biotic factors in an upperintertidal community: dipteran larvae grazing on algae. Ecology 62,1536–1547.

Santelices, B., 1990. Patterns of organization of intertidal and shallowsubtidal vegetation in wave exposed habitats in Central Chile.Hydrobiologia 192, 35–57.

Santelices, B., 2004. Ecological comparisons among aclonal (unitary),clonal and coalescing macroalgae. J. Exp. Mar. Biol. Ecol. 300,31–64.

Santelices, B., Vásquez, J., Meneses, I., 1986. Patrones de distribuciónydietas de un gremio de moluscos herbívoros en habitatsintermareales expuestos de Chile central. In: Santelices, Bernabé(Ed.), En: Monografías Biológicas No. 4. Simposio Internacional.Usos y funciones de las algas marinas bentónicas, pp. 147–171.

Santelices, B., Correa, J.A., Hormazábal, M., Flores, V., 2003. Contactresponses between spores and sporelings of different species,karyological phases and cystocarps of coalescing Rhodophyta.Mar. Biol. 143, 381–392.

Sousa, W.P., 1979. Experimental investigations of disturbance andecological succession in a rocky intertidal algal community. Ecol.Monogr. 49, 227–254.

Steneck, R.S., Dethier, M.N., 1994. A functional group approach to thestructure of algal dominated communities. Oikos 69, 476–498.


Steneck, R.S., Watling, L., 1982. Feeding capabilities and limitation ofherbivore molluscs: A functional group approach. Mar. Biol. 68,299–319.

Stimson, J., 1970. Territorial behaviour of the owl limpet Lottiagigantea. Ecology 51, 113–118.

Thompson, R.G., Tobins, M.L., Hawkins, S.J., Norton, T.A., 1999.Problems in extraction and Spectrophotometric determination ofchlorophyll from epilithic microbial biofilm: towards a standardmethod. J. Mar. Biol. Assoc. UK 98, 25–36.

Underwood, A.J., 1984. The vertical distribution and seasonalabundance of intertidal microalgae on a rocky shore in NewSouth Wales. J. Exp. Mar. Biol. Ecol. 78, 199–220.

Underwood, A.J., 1989. Design and analysis of field experiments oncompetitive interactions affecting behaviour of intertdal animals.In: Chelazzi, G., Vannini, M. (Eds.), Behavioural adaptation tointertidal life. Plenum press, New York, pp. 333–358.

Underwood, A.J., Jernakoff, P., 1981. Effects of interactions betweenalgae and grazing gastropods on the structure of a low-shoreintertidal algal Community. Oecologia 48, 221–233.

Vinueza, R.L., Branch, G.M., Branch, M.L., Bustamante, R.H., 2006.Top-down herbivory and bottom-up El Niño effects on Galapagosrocky-shore communities. Ecol. Monogr. 76, 111–131.

VonEnde,C.N., 1993.RepeatedMeasuresAnalysis: growth andother timedependent measures. In: Schneider, S.M., Gurevitch, J. (Eds.), Designand analysis of ecological experiments. Chapman & Hall, New York,pp. 113–137.

Wieters, E., 2005. Upwelling control of positive interactions overmesoscales: a new link between bottom-up and top-downprocesses on rocky shores. Mar. Ecol. Prog. Ser. 301, 43–54.

Wieters, E.A., Kaplan, D.M., Navarrete, S.A., Sotomayor, A., Largier,J., Nielsen, K.J., Véliz, F., 2003. Alongshore and temporalvariability in chlorophyll-a concentration in Chilean nearshorewaters. Mar. Ecol. Prog. Ser. 249, 93–105.

Williams, G.A., Davies, M.S., Nagarkar, S., 2000. Primary successionon a seasonal tropical rocky shore: the relative roles of spatialheterogeneity and herbivory. Mar. Ecol. Prog. Ser. 203, 81–94.

Wootton, J.T., Power, M.E., Paine, R.T., Pfister, C.A., 1996. Effects ofproductivity, consumers, competitors, and El Niño events on foodchain patterns in a rocky intertidal community. Proc. Natl. Acad.Sci. U. S. A. 93, 13855–13858.

Worm, B., Lotze, H.K., Hillebrand, H., Sommer, U., 2002. Consumerversus resources control of species diversity and ecosystemfunctioning. Nature 417, 848–851.

effects of chiton granosus (frembly, 1827) and other molluscan grazers on algal succession in wave...

Documents