Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

Contents lists available at ScienceDirect

Journal of Experimental Marine Biology and Ecology

j ourna l homepage: www.e lsev ie r.com/ locate / jembe

Invertebrate dispersal and habitat heterogeneity: Expression of biological traits in aseagrass landscape

Christoffer Boström ⁎, Anna Törnroos, Erik BonsdorffÅbo Akademi University, Environmental and Marine Biology, Artillerigatan 6, FIN - 20520 Turku, Finland

⁎ Corresponding author. Tel.: +358 50 3071814.E-mail address: cbostrom@abo.fi (C. Boström).

0022-0981/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.jembe.2010.05.008

a b s t r a c t

a r t i c l e i n f o

Article history:Received 1 October 2009Received in revised form 6 May 2010Accepted 14 May 2010

Keywords:Biological traitsDispersalPatch sizePhysical disturbancePost-settlementRecruitmentSeagrass landscape

Seagrass meadows harbour diverse faunal assemblages, but the relative importance of landscapes attributes,settlement processes and biological traits of individual species for recruitment patterns is poorly understood.To quantify the influence of habitat heterogeneity on larval, juvenile and adult post-larval dispersal,invertebrates (N125 µm) were collected in benthic settlement traps at five occasions (June–August) in threehabitats; continuous seagrass, seagrass patches and bare sand. The study was carried out by SCUBA diving ina subtidal (2.5 m depth) seagrass landscape dominated by Zostera marina L. in the Baltic Sea. Traps collectedtotally N30 taxa, with non-significant effects of habitat on species richness and total abundance. Totalnumber of invertebrates exhibited strong temporal variability, probably driven by wind-induced bedloadand water column transport. Surprisingly, traps located in small (b12 m2) patches contained on averagealmost twice as many individuals as traps located in the continuous vegetation. Dominating taxa such asnematodes, copepods, and oligochaetes were found in similar densities across the landscape. In contrast,location within the landscape had strong effects on bivalve settlement and redistribution patterns, resultingin significantly lower densities in continuous vegetation compared to patches and bare sand. A biologicaltrait analysis indicated that semi-mobile taxa with annual protracted direct development, and short-distancedispersal are probably of higher importance for the community assembly process in this environment thanlong-distance larval dispersal. Results suggest that seagrass landscapes are highly dynamic environments,characterized by time and species-specific effects of landscape attributes on animal dispersal andrecruitment. A conceptual model illustrating interactions between settling larvae and landscape hetero-geneity is presented.

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Organism–environment relationships are usually studied at thefine-scale in relatively homogenous environments, preventing pre-dictions of patterns and processes operating over broad spatial scalesin heterogeneous landscapes. To consider broad-scale spatial habitatpatterns is important and distinguishes the seascape approach fromtraditional marine ecological studies. Consequently, seascape ecologyhas developed as a valuable tool to encompass the complexheterogeneity that influences species–environment relationships inshallow water benthic ecosystems (Boström et al., 2006; Hincheyet al., 2008; Zajac, 2008). On shallow soft bottoms, scale, variabilityand spatial patterns have been acknowledged (Bergström et al., 2002;Hewitt et al., 2008), and broad-scale processes recognised in larvalrecruitment (Todd, 1998). However, studies extending the under-standing of these processes beyond single-species populations, to

whole assemblages and communities are rare (Menge and Suther-land, 1987; Kinlan and Gaines, 2003).

On soft substrates, dispersal patterns of marine invertebrates areknown to be highly variable in space and time (Commito and Tita,2002; Munroe and McKinley, 2007, Bowen and Hunt, 2009; Valankoet al., 2010). The variability in density of larvae, newly settled andredistributing individuals is caused by complex interactions betweenhydrodynamics, habitat structure, habitat configuration, predationand species-specific traits (Armonies, 1994; Hunt and Scheibling,1997; Bologna and Heck, 2000; Commito and Tita, 2002). Softsediment benthic systems are in this perspective impeded bydifficulties in both logistic and practical sampling problems as wellas quantification of settlement compared to hard substrate systems,but methods to tackle these issues are emerging (e.g. Hewitt et al.,2007; Valanko et al., 2010).

Settlement, i.e. change in larval activity to a benthic life stage, on softsediments differs from hard substrata in that the attachment to thesubstrate is not permanent (Commito and Tita, 2002), and ongoingredistribution is a common feature, especially in tidal environments(Bowen and Hunt, 2009). The term recruitment generally refers tosettled individuals that survive an arbitrary time-period to adulthood

107C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

and contribute to the assemblage (Connell, 1985). In this context,life-history and functional traits such as reproductive type, feedinghabits and larval type have been identified as important factorsinfluencing dispersal, settlement and community development(Tamaki, 1987; Armonies, 1994; Commito and Tita, 2002; Bremneret al., 2003a, b).

Seagrass meadows sustain important ecological services in coastalwaters but show increasing loss rates due to human activities(Waycott et al., 2009). Due to human impact, and also due to naturalprocesses such as physical disturbance and grazing, continuousmeadows are fragmented into heterogeneous landscape-mosaicsconsisting of smaller, isolated patches (Boström et al., 2006). Howsuch continuous changes in habitat structure influence invertebratedispersal and community assembly is poorly understood. In partic-ular, the relative contribution of habitat characteristics vs. speciestraits (Bremner et al., 2003a, b) for the structure of benthiccommunities has not been contrasted with recruitment studies.

Both plant and landscape attributes influence settlement patterns.Thus, structurally complex plant assemblagesmay trap organismsmoreefficiently than structurally simple leaf canopies (Hovel and Fonseca,2005; BoströmandBonsdorff, 2000). In contrast, according to the “settleand stay”-hypothesis (Bell and Westoby, 1986), settling larvae aredistributed patchily, and do not discriminate among seagrass patches ofdiffering complexity, suggesting that the physical location of the patchin relation to the larval source pool is critical, and more important thancomplexity itself (Sogard, 1989). Some studies further suggest that dueto increased perimeter:area (P:A) ratios in small vegetation patches, theprobability of encounter rates of passively dispersing fauna should behigher in patchy landscapes (Bologna and Heck, 2000).

The effects of habitat complexity on flow dynamics have beenexperimentally demonstrated in flumes (Butman et al., 1988), andresponses of single species to seagrass landscape structure (Bolognaand Heck, 2000) and other biogenic structures (Eggleston et al., 1999;Commito et al., 2005) have been studied extensively. However, thereis a scarcity of community level studies focusing on how seagrasslandscape attributes influence dispersal and diversity of the wholemacrofaunal assemblage. In particular, it is unclear how habitatheterogeneity interacts with temporally variable dispersal andrecruitment patterns of infauna.

To quantify the influence of habitat heterogeneity on larval,juvenile and adult post-larval dispersal, we established permanentsettlement traps in three parts of a high-energy seagrass landscape;on bare sand, in small isolated seagrass patches and in continuousvegetation, and sampled the traps at five occasions during thesummer season. Based on previous studies on physical (Gambiet al., 1990) and biological, i.e. settlement (Orth, 1992; Bologna andHeck, 2000) processes in seagrass ecosystems, and unstructured soft-sediment environments (Valanko et al., 2010), we hypothesized thataccumulation of individuals would be greatest in patches and lowest inbare sand, while the meadow should exhibit intermediate densities.While species traits such as feeding mode and mobility have beenacknowledged as an important driver of local scale succession dynamics(Zajac et al., 1998), these issues have not been specifically addressedwithin an experimental framework in heterogeneous seascapes, but seeHewitt et al. (2008) for a descriptive approach. Due to the obviousphysical differences between these habitats in seagrass–sand mosaics,we further hypothesized that the invertebrate assemblages in thesehabitats would be characterized by highly different biological traits interms of e.g. mobility, reproduction and feeding habit.

The following specific questions were posed:

(1) To what extent are larval and post-larval abundance, diversityand dispersal patterns influenced by seagrass landscapeattributes?

(2) What biological traits characterize the faunal assemblages andare traits expressed differently across the landscape?

2. Material and methods

2.1. Study area and site characteristics

The study site is located at the Hanko Peninsula at the entrance ofthe Gulf of Finland, in the northern Baltic Sea (Fig. 1). This area ischaracterized by extensive (10–20 ha) seagrass meadows dominatedby Zostera marina interspersed with Potamogeton pectinatus (Boströmet al., 2002). The study was conducted in a subtidal (2–3 m depth)seagrass landscape (Ryssholmen: 59°60′N, 23°05′E) exposed tosoutherly winds (90–180°), which prevail during the summermonths(June–August). Data on daily wind speed and direction during thestudy period was obtained from a nearby (b10 km, Russarö) weatherstation operated by the Finnish Meteorological Institute. The studiedseagrass landscape consists of a∼7 ha Zosterameadow characterized byelongated bare sand areas in the interior parts, and more continuousvegetation towards the edges, while isolated seagrass patches are foundin the more exposed areas south of the meadow edge (Fig. 1). Thepatches are highly variable in terms of size (5–200 m2) and shape, andform a mosaic landscape consisting of both circular and elongatedseagrass patches (perimeter:area ratios 0.5–2.7).

2.2. Invertebrate traps and field sampling

To investigate potential spatial and temporal differences in larvaland post-larval abundance, diversity and dispersal patterns across theseagrass landscape, we established clusters of three permanent trapsin three distinct microhabitats in the seagrass landscape: in bare sand(hereafter “Sand”), in small isolated patches (hereafter “Patch”), andin the continuous meadow (hereafter “Meadow”). This design wasreplicated (n=2) spatially (Fig. 1), and the traps clustered in eachhabitat were treated as subsamples and averaged in the statisticalanalysis (see below). To avoid confounding effects of varying depths,the bare sand traps were located 20 m from the patches towards theopen sea. As it was impossible to locate patches of exactly the samesize and shape with similar degree of isolation (i.e. minimum of 3 mfrom the nearest neighbouring patch), traps were established in twodifferent sized patches, namely 4 and 11 m2, but with similarperimeter to area ratios, 0.62 and 0.56, respectively. These patchsizes represent the smallest at the study site. The patcheswere locatedin the outermost zone of the patchy area about 60 m from the edge ofthe continuous meadow (Fig. 1). As the most interior parts of themeadow were deeper and contained bare sand areas with occasion-ally dense accumulations of drifting algal mats, and as the goal withthe designwas not to compare the intensity of edge effects in differentsized patches, traps in the meadow were positioned in continuousvegetation at least 10 m from the southern meadow edge, andconsidered to represent core area of continuous vegetation (Bolognaand Heck, 2000), while the other vegetation traps were placed in thecentre of the isolated patches 1–1.5 m from the patch edge.

A trap was constructed of a pvc tube (diameter 10 cm, length50 cm) equipped with a 350 ml removable glass jar (aspect ratio of2.2:1). The opening of the jar was kept at the same level as the upperend of the tube bymeans of a bolt below the jar. To avoid trapping andclogging of drift algae, but allow trapping of invertebrates, traps werecovered with a 10 mm mesh. This was a trade-off, as the meshprobably reduced trap efficiency. To ensure survival of organisms, thejar contained 1 dl of sieved (b500 µm) azoic sand. The tubescontaining the traps were pushed into the sediment to a depth ofabout 20 cm, leaving the opening of the trap at the same height(30 cm) as the upper parts of the leaf canopy. Traps in bare sand werepositioned at the same height as the vegetation traps. Protruding trapsgenerally capture less benthic invertebrates compared to bottomtraps (Smith and Brumsickle, 1989; Valanko et al., 2010). However,due to the high exposure, mobile sandy sediment and longdeployment times, bedload traps flush with the sediment surface

Fig. 1. Location of the study area and aerial photograph of the seagrass landscape. The approximate location of traps (white dots) in the meadow, patches and bare sand is shown.Photograph: Lentokuva Vallas Oy.

108 C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

could not be used. Thus our traps were primarily aimed at collectingorganisms displaying active behavioural emergence from the sedi-ment surface, larvae settling vertically down from the water column,and resuspended organisms. Clusters of three traps were placed 1 mapart in a triangular pattern in each habitat. The traps were sampledby SCUBA diving at five occasions during the most productive period

in the study area, i.e. from the beginning of June (June 3) to the end ofAugust (August 22). Due to varying weather conditions and logisticalconstraints, the field sampling procedure could not be accomplishedat exact time intervals, thus open-trap capture times varied slightly(14–20 days). At each sampling occasion, jars were carefully closed,retrieved and replaced by new jars with azoic sediment.

Table 1Biological trait variables, categories and corresponding labels selected to indicatefunctional properties of invertebrates dispersing in a seagrass landscape.

Biological trait Trait categories Labels

Mean (adult) size 0.1–1 mm VS1–5 mm S5 mm–1 cm SM

Larval type Planktotrophic PlanktoLecitotrophic LecitoDirect development Direct

Reproductive frequency Semelparous SemelAnnual episodic Annu_epiAnnual protracted Annu_prot

Propagule dispersal range b0.1 m VL (very low)0.1–1 m Low1–10 m LM (low–medium)10–100 m Medium100–1000 m ME (medium–extensive)1000–10000 m ExtensiveN10000 m VE (very extensive)

Environmental position Infauna (N5 cm) Inf_deepInfauna (2–5 cm) Inf_middleInfauna (top 2 cm) Inf_topEpibenthic EpiBentho-pelagic Bentho_pel

Feeding habit Detritivore DetritivoreOmnivore OmnivoreHerbivore HerbivoreCarnivore CarnivoreParasite Parasite

Mobility (adult) Mobile MobSemi-mobile Semi_mob

Movement method (adult) Swimmer SwimmerRafter/drifter Raft_driftCrawler CrawlerBurrower Burrower

Dispersal type (adult) Resident ResidentLocal LocalLong-distance Long_dist

109C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

To test for potential differences in Z. marina shoot density betweenthe meadow and the patches, shoots were counted around theposition of the traps using a 25×25 cm frame (n=5/habitat).

2.3. Sample processing

In the laboratory, samples were fixed with 70% alcohol and stainedwith Bengal Rose. The animals were separated from the sediment byrinsing. Thus, a sample was completely resuspended for 15 min in atransparent pvc cylinder (diameter 9 cm, length 50 cm) with anoverflow. A tube was inserted to the bottom of the cylinder to create acurrent sufficiently strong to remove all organic particles andorganisms, but not the sediment. Analysis of the remaining sedimentof 2 test samples revealed that this method captured N95% of allanimals. All animals retained on a 125 µm sieve were enumerated andidentified under a dissection microscope (10 times magnification) tothe lowest taxonomic level possible.

2.4. Numerical analysis

To analyse the influence of habitat (Sand, Patch, Meadow) overtime (5 occasions) on total abundance and number of species we useda univariate repeated-measures ANOVA (rmANOVA). In this analysis,between-subject effects (Habitat) test the treatment effect averagedover time, while within-subject effects (Time, Time×Habitat)represent a test of how the treatment influences changes in a variableover time. Because of too few bivalves on some dates, the abundancesof the two dominating bivalve species were subjected to a similaranalysis, except for using only 3 (June 18, July 8, and August 6) and 2

(July 8 and August 6) dates for Macoma balthica and Cerastodermaglaucum, respectively.

Prior to analysis, dependent variables were tested to meetassumptions for parametric testing (normality: Kolmogorov–Smir-onov test; homoscedasticity: Cochran's test) and square root(abundance) or log transformed (individual species) when necessary.Mauchly's test of sphericity was used to assess if the covariancematrix met the sphericity assumption (von Ende, 1993). When thesphericity assumption was violated, we used Huynh–Feldt epsilonadjusted F values and probabilities (Huynh and Feldt, 1976). Aposteriori comparisons (alpha 0.05) for between-subject effects weremade using Scheffe's test, while pairwise comparisons of within-subject effects were made using Bonferroni adjusted multiplecomparisons. Analyses were conducted in SPSS and SuperANOVA(Abacus Concepts). All means are reported±1SE.

To quantify and compare possible differences in the temporalturnover of species in different parts of the seagrass landscape, weused Whittaker (1960) measure (βW). βW=S /α−1, where S=totalnumber of species recorded in the system and α=the average samplediversity, where each sample is of standard size and diversity ismeasured as number of species (Magurran, 1991). When used tocompute βW between pairs of samples, values range from 0 (completesimilarity, low turnover) to 1 (no overlap in species composition, highturnover).

2.5. Biological Trait Analyis (BTA)

To compare the assemblage pattern in different habitats from afunctional perspective, we assigned a number of biological traits to theten most abundant taxa. Nine biological traits, sub-divided into 35categories (Table 1), were selected to indicate functional propertiesrelevant for a better understanding of community dynamics and plant–animal interactions (McGill et al., 2006; Frid et al., 2008). Trait categoriesweremodified to apply for benthos in the northern Baltic Sea (Bremneret al., 2003a,b; Valanko et al., 2010), and are based on both scientificliterature as well as expert consultations (A. Törnroos, unpubl. data).

To quantify functional properties of trapped invertebrates, weconducted a biological trait analysis (BTA; Bremner et al., 2003a) andassigned species to traits using a “fuzzy coding” procedure (Charvetet al., 1998) and applied a weighing procedure incorporating theabundance of species (Bremner et al., 2003b; Hewitt et al., 2008). Thedegree towhich a species exhibited a specific trait categorywas scaledbetween 0 and 3 and the category scores standardized to 1 within atrait. By multiplying the trait category scores by the mean speciesabundances in each habitat and date, a trait matrix was constructed.Abundances were fourth-root transformed in order to down-weightthe influence of very abundant taxa. Based on the matrix, relative (%)trait expression was calculated within each habitat over all dates. Toanalyse trait composition in each habitat over time a FuzzyCorrespondence Analysis (FCA, Chevenet et al., 1994) was conducted(Bremner et al., 2003a). This analysis is a powerful tool in detectingspatial and temporal changes in trait expression within and betweenassemblages (Bremner et al., 2006). The abundance-weighted scoresare ordinated on a two-dimensional plane based on Euclideandistances of the relative frequencies of trait scores in each sample(Bremner et al., 2003b). For this analysis we used the two replicatesfrom each habitat and date (n=2/time/habitat). The FCA was done inthe R environment (R Developmental Core Team, 2009).

3. Results

3.1. Density and diversity of trap fauna

There was no significant difference (t-test; p=0.48) in shootdensity between the patches (average: 544±60 shoots m−2) and themeadow (average: 611±70 shoots m−2), indicating that possible

Table 2Species list of fauna collected in the settlement traps during the study period (June–August) in bare sand, small patches and in the continuous meadow. + or− indicates ifa species was present or absent, respectively, in either of the two transects. Numberswithin brackets refer to the number of sampling dates a species was found, e.g. 3/5indicates that a species was encountered on 3 out of 5 sampling occasions.

Class Species Sand Patch Meadow

Hydrozoa Hydra sp. + (4/5) + (4/5) + (4/5)Protohydra sp. + (3/5) + (4/5) + (4/5)

Turbellaria + (5/5) + (5/5) + (5/5)Nemertinea Prostoma obscurum + (4/5) + (3/5) + (2/5)Nematoda + (5/5) + (5/5) + (5/5)Priapulida Halicryptus spinulosus + (4/5) + (4/5) + (4/5)Polychaeta Nereis diversicolor − + (2/5) + (2/5)

Polychaeta larvae + (2/5) + (3/5) + (2/5)Oligochaeta + (5/5) + (5/5) + (5/5)Hirundinea Piscicola geometra + (4/5) − + (5/5)Mollusca Mytilus edulis + (3/5) + (2/5) + (1/5)

Cerastoderma glaucum + (5/5) + (5/5) + (5/5)Macoma balthica + (5/5) + (5/5) + (5/5)Mya arenaria + (2/5) + (2/5) −Theodoxus fluviatilis + (2/5) + (3/5) + (4/5)Potamopyrgus antipodarium + (1/5) + (1/5) + (1/5)Hydrobia spp. + (5/5) + (5/5) + (5/5)Limapontia capitata + (3/5) + (3/5) −Embletonia pallida + (1/5) + (1/5) + (2/5)

Bryozoa Electra crustulenta + (1/5) − + (1/5)Arachnida Hydrachnida + (5/5) + (5/5) + (5/5)Crustacea Cladocera + (5/5) + (5/5) + (5/5)

Cercopagis pengoi − − + (1/5)Ostracoda + (5/5) + (5/5) + (5/5)Copepoda + (5/5) + (5/5) + (5/5)Balanus improvisus + (5/5) + (5/5) + (5/5)Idotea chelipes − − + (1/5)Idotea balthica − + (2/5) + (2/5)Jaera albifrons + (4/5) + (4/5) + (5/5)Gammarus spp. + (5/5) + (5/5) + (5/5)Monoporeia affinis + (1/5) + (1/5) + (1/5)Corophium volutator + (3/5) + (3/5) + (4/5)

Insecta Chironomidae + (5/5) + (5/5) + (5/5)Podura aquatica + (4/5) + (2/5) + (2/5)

Pisces Pomatoschistus minutus − + (1/5) + (1/5)Varia + (4/5) + (3/5) + (4/5)Total 36 31 32 34

110 C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

faunal differences were not due to different within-patch character-istics (habitat complexity) between the vegetated habitats. Thecomposition and total number of species collected in the traps aresummarized in Table 2. The average (mean values for the whole

Fig. 2. Number of individuals (a) and species (b) collected

sampling period) patterns in density and diversity of the capturedfauna exhibited relatively high densities in sand and patches, lowerdensities in the meadow, and equal number of species in all habitats(Fig. 2). The total number of captured organisms showed strongtemporal variability, and two pronounced peaks (July 8 and August 6)were recorded during the study period (Fig. 2a). Consequently, thermANOVA revealed a significant effect of Time (F4,12=74.2,p=0.0001) on total abundance (Table 3). Traps located in patchescontained on average almost twice as many individuals (1595±272)as traps in the meadow (891±121), but variability was substantial,and resulted in a non-significant (F2,3=5.2, p=0.1) effect of Habitaton total abundance (Table 3). However, low statistical power (n=2)might mask potential differences. During the highest settlement peak(July 8) traps located in patches collected most individuals, followedby the sand habitat and the meadow (Fig. 2). However, the influenceof habitat within the seagrass landscape was dependent on samplingdate, resulting in a significant interaction between these factors(Table 3). These interactions demonstrate the nature of the variabilityin invertebrate abundances across habitats in this seagrass landscape.For example, during the first peak in abundance (July 8) the patchesscored the highest (∼4000 individuals), followed by the sand habitat(∼3000 individuals) and the meadow (∼2000 individuals), whileduring the second, less pronounced peak (August 6), the sand habitatexhibited the highest abundances, followed by the patch and themeadow. On all other dates, faunal numbers were similar acrosshabitats. Interestingly, the continuous meadow habitat showedconsequently the lowest invertebrate densities.

The diversity of organisms did not differ among habitats within thelandscape (F2,3=0.05, p=0.9, Table 3), but significant temporalvariability in species richness was recorded (F4,12=6.4, p=0.005,Table 3, Fig. 2b). In particular, there were on average 9 species (taxa)in all habitats sampled, with the highest number of species coincidingwith the two peaks in total density (Fig. 2). In contrast to densitypatterns, the number of species showed no significant interactionbetween time and habitat (Table 3), suggesting a similar temporalpattern regardless of habitat in the landscape.

3.2. Species composition and bivalve settlement patterns

The compositions of the faunal assemblages were remarkablysimilar in time and space. Nematodes, copepods and oligochaetesranked among the top four taxa regardless of habitat within the

in the traps during the study period (June–August).

Table 3Repeated-measures ANOVA on impacts of habitat total abundance and species richnessof invertebrates in settling traps sampled from June to August (five sampling dates).Within-subject p values refer to Hyundh–Feldt corrected probabilities (see Materialsand methods).

Source Abundance Number of species

df MS F p MS F p

Between-subjectsHabitat 2 173.643 5.240 0.1050 0.233 0.050 0.9524Error 3 33.137 4.704Within-subjectsTime 4 1197.759 74.232 0.0001 12.154 6.372 0.0055Time×Habitat 8 52.093 3.228 0.0333 1.229 0.644 0.7287Error 12 16.135 1.907

111C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

seagrass landscape (Fig. 3). These taxa, along with ostracods(meadow), the bivalve M. balthica (patch), and the lagoon cockleC. glaucum (sand) constituted 85%, 84% and 81%, respectively of thetotal abundance in the three habitats sampled. In separate rmANO-VAs, the dominating groups showed non-significant responses tohabitat (nematodes: p=0.10, copepods: p=0.41 and oligochaetes:p=0.88), but significant temporal variability (Time: all taxap=0.0001). Multiple comparisons (p=0.05) for taxa exhibiting anon-significant Habitat×Time interaction, revealed that nematode

Fig. 3. Species composition and relative abundance of the ten numerically most important tstudy period (June–August). Note the decreasing importance of bivalves (bold) in sand to m

abundance peaked on date 2, and differed significantly from dates 3, 4and 5. Copepods peaked on date 1, and abundance was significantlyhigher than on dates 3, 4, and 5, while date 2 exhibited significantlyhigher abundance than date 4.

The contribution of captured bivalves (M. balthica and C. glaucum)to the total abundance varied across the scale of the studiedlandscape. The contribution of bivalves to the total abundance wasmost significant in the sand habitat (C. glaucum andM. balthica 10.8%and 5.7%, respectively, Fig. 3), while traps located in the patchestrapped fewer (M. balthica 5.6% and C. glaucum 3.6%) settlingbivalves (Fig. 3). The lowest number of bivalves was recorded in themeadow (b2%, Fig. 3). Habitat within the seagrass landscape hadstrong effects on bivalve settlement and redistribution. In particular,continuous vegetation appeared to inhibit the settlement of bothM. balthica and C. glaucum, while traps located in bare sand and inthe patches exhibited constantly much higher densities (Fig. 4a, b).Thus, the influence of habitat was highly significant (F2,3=282.9,p=0.0004) and marginally significant (F2,3=9.2, p=0.052), forM. balthica and C. glaucum, respectively (Table 4). Pair-wise com-parisons (Scheffe's test) for M. balthica revealed non-significantdifferences between bare sand and the patches (p=0.46), whileboth these habitats showed significantly higher abundances thanthe meadow (Meadow vs. Patch: p=0.007, Meadow vs. Sandp=0.006).

axa collected in the traps in bare sand (a), patches (b) and the meadow (c) during theeadow habitats.

Fig. 4. Density of settled and resuspended bivalves;Macoma balthica (a), Cerastoderma glaucum (b) in traps located in sand, patches and the meadow during the study period (June–August) in relation to maximumwind speed (c) at the study site. The wind graph represents wind speeds from all directions, but peaks from directions directly influencing the studysite, i.e. SW–SE are indicated with circles. In panel (c), dashed arrows indicate sampling dates.

112 C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

Strong southerly winds appeared to increase transport of bivalves.As a qualitative indication of thismechanism, the first settlement peakof C. glaucum (July 8) was preceded by two maxima in southerlywinds (19.8 and 17.3 m s−1 on June 22 and July 4, respectively,Fig. 4c). Similarly, the abundance peak on August 6 was againpreceded by strong southerly winds, e.g. 14.7 m s−1 on July 23(Fig. 4c). The size of the captured bivalves increased over time (Fig. 5).Thus, bivalves had reached a relatively large size on August 6 (750 and1500 µm forM. balthica and C. glaucum, respectively), suggesting thatthe contribution of pelagic settlers diminished towards the end of theseason, while the importance of wind-driven resuspension andsecondary settlement increased especially in the patches and onbare sand. The smaller mean size of both bivalve species in the

Table 4Whittaker's measure (βW) of species turnover between sampling occasions in trapslocated in different habitats in the seagrass landscape (mean values: n=3). Valuesrange from 0 to 1, with low values indicating high similarity or low turnover, whilevalues close to 1 indicate no overlap in species composition or high turnover.

Period Western transect Eastern transect

Sand Patch Meadow Sand Patch Meadow

18.6–8.7 0.06 0.06 0.00 0.05 0.16 0.128.7–22.7 0.06 0.06 0.03 0.14 0.13 0.1222.7–6.8 0.00 0.12 0.06 0.14 0.13 0.106.8–22.8 0.12 0.09 0.09 0.12 0.10 0.05Total period (18.6–22.8) 0.12 0.16 0.10 0.12 0.16 0.14

meadow on August 6 (Fig. 5) indicates lesser importance ofresuspension in continuous vegetation, where also the physicaltrapping of algae may prevent secondary settlement.

The consistently very low numbers of Whittaker's βW in the trapdata, i.e. 0.06–0.16, indicated high similarity or only minor exchangeof species between sampling occasions and habitats (Table 5), i.e. thespecies pool of the entire landscape is relatively constant throughoutthe season.

3.3. Expression and importance of biological traits

The composition of traits showed high similarity between thehabitats studied, and clearly reflected the dominance patterns of theindividual species (Fig. 3). Thus all species assemblages, regardless ofhabitat within the seagrass landscape, were characterized byorganisms expressing a semi-mobile means of mobility by crawling,a detrivorous feeding habit and direct development (Fig. 6). The FCAhighlighted three groupings of samples (Fig. 7a), and the two FCAaxes together accounted for 72.5% of the variability in traitcomposition between the dates, i.e. 47.7% and 24.8% respectively(Fig. 7b). Group 1 represented the sand and patch samples from dates1 and 2, Group 2 corresponded to all sand and patch samples fromdate 4, while the Group 3 contained only samples from the meadowon date 5.

Group 1 was distinguished by trait categories located to the upperleft in Fig. 7b, corresponding to burrowing infauna with a semelpar-ous reproductive frequency and very low propagule dispersal range,

Fig. 5. Average size distribution of bivalves (a)Macomabalthica, (b) Cerastoderma glaucumcollected in traps located in sand, patches and the meadow. Mean sizes are based on theentire data set.

113C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

while Group 2 corresponded to trait categories located to the lowerleft in the ordination (Fig. 7b), e.g. infaunal species with annualepisodic reproductive frequency, burrowing movement method and adetritivorous feeding habit. Trait categories distinguishing Group 3was less clear, however features such as epibenthic sedimentlocation, annual protracted reproduction frequency and propaguledispersal distances ranging from low to extensive could be related tothis group (Fig. 7b).

4. Discussion

This study provides information about spatial (one to tens ofmeters) and temporal (months) scales over which post-settlementdispersal occurs and how habitat configuration interacts with thisprocess. To our knowledge this study also represents one of the first

Table 5Repeated-measures ANOVA of impacts of habitat on abundance of Baltic clams(Macoma balthica) and lagoon cockles (Cerastoderma glaucum) in settling trapssampled from June to August. Numbers of sampling dates were three and two for M.balthica and C. glaucum, respectively. Within-subject p values for C. glaucum refer toHyundh–Feldt corrected probabilities (see Materials and methods).

Source Macoma balthica Cerastoderma glaucum

df MS F p df MS F p

Between-subjectsHabitat 2 2.118 282.989 0.0004 2 1.118 9.236 0.0522Error 3 0.007 3 0.121Within-subjectsTime 2 0.016 0.224 0.8056 1 0.319 12.271 0.0394Time×Habitat 4 0.454 6.359 0.0238 2 0.106 4.076 0.1395Error 6 0.071 3 0.026

attempts to quantify how biological traits are expressed in relationto seagrass landscape features. Such data has recently beenidentified as a crucial gap in knowledge (Jennings and Hunt,2009), and in seagrass systems restricted to more traditionalapproaches in terms of functional groups/guilds (Hily and Bouteille,1999; Boström et al. 2006). The results suggest that this high-energy seagrass–sand mosaic is interlinked by a considerable flowof invertebrates (Fig. 8), resulting in a surprisingly high similaritybetween microhabitats and assemblages. In accordance with aprevious work from this region (Boström and Bonsdorff, 2000) andelsewhere (Palmer, 1988; Commito and Tita, 2002), resuspensionand subsequent redistribution of species with direct developmentappear to be of key importance. The planktotrophic pool of larvae isprimarily found outside the seagrass meadow, and only expressedto a minor extent in the trapped fauna, while the pool of directdeveloping larvae is present in all parts of the landscape, andexpressed to higher extent (Figs. 6 and 8). The seagrass landscapewas also characterized by a remarkably low species turnover(Table 5), indicating that both seagrass patches and the continuousmeadow receive a similar supply of species from a local species poolthroughout the summer season. Contrary to our predictions, trapslocated in bare sand collected equal, and occasionally higherdensities of invertebrates compared to traps positioned in themeadow (Fig. 2). Particularly the habitat in which settlement trapswhere located had a strong influence on the settlement pattern ofM. balthica, with lower densities in the meadow and similarpatterns were observed for C. glaucum. In accordance with ourprediction, traps located in the small patches collected moreindividuals than the traps located in the meadow.

4.1. Seagrass landscape configuration and faunal distribution

The significant influence of spatial arrangement and size ofhabitat patches for ecological processes, i.e. landscape/seascapeecology, has been recognised in the study of coastal environments(Hinchey et al., 2008), particularly in seagrass ecosystems (Robbinsand Bell, 1994). In the context of larval recruitment to our seagrasslandscape, at least three aspects are relevant to consider, namely(1) edge effects, (2) the influence of patch size and (3) theimportance of patch shape and orientation relative to the prevailingwater flow.

Edges of seagrass beds may exhibit higher abundances of settlers,and higher animal density and production than interior parts(Bologna and Heck, 2002; Warry et al., 2009). Given the lowerresuspension rates in seagrass (Gambi et al. 1990, Hasegawa et al.,2008), especially in high-energy environments (Gacia et al., 1999) it isreasonable to suggest that the bulk of captured larvae, post larvae andjuvenile recruits of macro-invertebrates were likely passively trans-ported from the open exposed sandy substrates outside the seagrassmeadow into the vegetated area. Consequently, the continuouslylower recruit densities in the meadow, especially after southerly windpeaks (Fig. 4) may demonstrate the presence of a “settlementshadow” (Orth, 1992). Thus, the highest densities of passivelydistributed larvae are trapped along the outermost meadow edge,and reduced flow rates towards the interior of the bed result inreduced larval settlement (Bologna and Heck, 2002). Given the factthat our traps were not positioned at the seagrass–sand interface, andthat the deceleration of flow occurs within the first 0.5 m into thevegetation (Peterson et al., 2004), reduced densities of settlers(especially bivalves) in the meadow may be explained by thismechanism.

This study also identified, from a northern Baltic perspective, ahigh total number (N30, Table 1) of invertebrate taxa, and similaraverage levels of species richness (9 species, Fig. 2b) regardless ofhabitat within the seagrass landscape, suggesting that animaldiversity in the traps well represents the natural communities in

Fig. 6. Trait composition and relative expression (%) of the ten most important trait categories in traps in bare sand (a), patches (b) and the meadow (c). Bars represent pooled datafrom the entire study period (June–August).

114 C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

the area (Boström et al., 2006). Seagrass patch shape andorientation is important especially in environments with highdegree of passive dispersal and active daily movement rates byadult epifauna (Tanner, 2003). Relatively higher abundance ofsettlers and resident fauna in small patches has frequently beenexplained by the higher encounter rates of fauna in small patchesrelative to larger patches, due to high and low P:A ratios,respectively (Eggleston et al., 1998). This mechanism might explainthe trend of occasionally high settling densities recorded in thepatches relative to the meadow, especially for bivalves (Figs. 2 and4). The limited empirical evidence for the interaction among patchshape, orientation and faunal movement in seagrass, suggests thatadult individuals, and also likely passively moving larvae, respondstrongly to patch shape and orientation, especially under highcurrent velocity (Tanner, 2003, Fig. 8). The high numbers and post-larval size (Fig. 5) of bivalves in traps located in bare sand implicatea major importance of bedload and water column transport formollusc dispersal in this seagrass landscape.

4.2. Interactions between species traits and landscape attributes

In marine systems, it is generally presumed that the continuoussupply of larvae connects and shapes populations in heterogeneoushabitats (Gaines et al., 2007). Although pelagic development andplanktotrophic larval types predominate in the marine realm, thespecies numerically most represented in our data have directdevelopment in the sediment. The dominance of direct developers

in soft bottom benthic recruitment studies is a well documentedphenomenon (Martel and Chia, 1991;Whitlatch et al., 1998; Commitoand Tita, 2002; Junkins et al., 2006; Johnson et al., 2007), andcorresponds well with a previous experimental work on exposedsandy bottoms in the study area, in which species reproducing viapelagic larvae only comprised about 30% of the whole settlingassemblage (Boström and Bonsdorff, 2000). It has been argued,however, that even a small group of individuals, a drifting egg mass oreven just one fertilized female with direct development may rapidlyestablish a population if the environment is suitable (Johannesson,1988). In addition, three additional biological traits appeared to beimportant functional features of the benthic fauna in this landscape,namely semi-mobility, a crawling movement method and detrivory(Fig. 6, Table 5).

Although settlement and recruitment patterns on disturbed and/orstructurally heterogeneous soft bottoms have shown to be verycomplex (Peterson, 1986; Smith and Brumsickle, 1989; Armonies,1994; Eggleston et al., 1999; Commito et al., 2005), our study showedthat species turnover was low (maximum Whittaker's βW=0.16,Table 5), and that the subsequent expression of trait composition wassurprisingly uniform across this heterogeneous landscape (Fig. 6). The10most expressed traits indicate amajor role of sporadic wind-drivendynamics for larval, post-larval and adult movement patterns(Valanko et al., 2010) and temporal trait differences were evident fortwo habitats: the bare sand and the patches. For these two habitats,three dates (June 18, July 8 and August 6) were distinguished from therest of the samples, forming two distinct groupings (Group 1 and

Fig. 7. Ordination of samples (dates) (a) along the first two axes in the Fuzzy Correspondence Analysis (FCA) of abundance-weighted biological traits, and (b) ordination of traitcategories with eigenvalue histogram. Samples highlighted by circles represent groupings of specific biological trait data. Group 1 represents all samples from the sand and patchhabitats on sampling dates 1 (June 18) and 2 (July 8), Group 2, all samples from sampling date 4 (August 6) in sand and patch, and Group 3 represents the twomeadow samples fromdate 5 (August 22). For variable labels, see Table 2.

115C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

Group 2, Fig. 7). This temporal pattern of traits was due to theincreased total abundances of fauna, which generated differences inthe scores (Hewitt et al., 2008). The trait composition of Groups 1 and2included the extremes of the trait category spectra, e.g. semelparousand annual episodic development, and very low and very extensivepropagule dispersal ranges. These two groups also showed differenttrait compositions that could, for Group 2 (patch and sand, August 6)be related to high numbers of mussels (M. balthica and C. glaucum),exhibiting traits such as planktotrophic larval typewith very extensivepropagule dispersal ranges. In comparison, Group 1 contained infaunalspecies with semelparous reproduction, e.g. traits expressed bynematodes (Fig. 7).

We recorded a seasonal pulse of species with planktotrophic larvae(M. balthica and C. glaucum) in the early summer, during a period of

about one month (Bonsdorff et al., 1995). This illustrates how the firstsets of settlers (≤300 μm size) establish from the water columnequally in all habitats, and later larger individuals (N500 μm size) areredistributed unevenly bywind-driven bedload transport between thehabitats (Fig. 5). This study further showed that juvenile and larvalstages of not only planktotrophic taxa, but particularly taxa with traitssuch as annual protracted direct development and short-distancepropagule dispersal (Nematoda, Oligochaeta and Turbellaria) are ofequal, if not of higher, importance for community assembly. Weconclude that in this high-energy seagrass–sand mosaic, it is not onlythe delivery of new individuals produced elsewhere that is important,but also the local retention and redistribution of short-distancedispersers, which may play a significant role for the seagrasscommunity assembly and structure.

Fig. 8. Conceptual model illustrating main directions and mechanisms influencing spatial dynamics of larval movements in a fragmented seagrass landscape characterized by highdegree of physical exposure. The same landscape is illustrated from above to depict landscape heterogeneity, and as a cross section to illustrate substrate heterogeneity provided byseagrass patches. Patches outside the continuousmeadow as well as the edge region of the continuousmeadow act as a larval filter reducing the number of settlers in interior parts ofthe continuous vegetation. Small patches with high edge:area ratios may initially exhibit high numbers of settlers due to increased encounter rates, but numbers decrease over timedue to continuous disturbance. Two distinct source pools are identified; a direct developing pool contributing to local assemblages via resuspension (relative importance illustratedby the size of brown, winding arrows), and a more remote pelagic pool of planktotrophic larvae which settle into the landscape directly. Once settled, planktotrophic larvae may, to avarying degree, enter the pool of resuspending animals, especially on unstructured sediment. The varying sizes of the direct developing larval pool/re-settling individuals illustratethe relative importance of resuspension across the landscape.

116 C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

Acknowledgements

This study was financed by the Academy of Finland (CB ProjectSEASCAPE 200689, decision number 80509), by the Foundation forÅbo Akademi University (EB) and by the Maj and Tor NesslingFoundation (AT). We thank J. Lindholm, S. Söderström and K.Tötterman for their help in the field and laboratory. TvärminneZoological Station is acknowledged for providing excellent facilitiesduring the field work.[ST]

References

Armonies, W., 1994. Drifting meio- and macrobenthic invertebrates on tidal flats inKönigshafen: a review. Helgol. Meeresun. 48, 299–320.

Bell, J.D., Westoby, M., 1986. Abundance of macrofauna in dense seagrass is due tohabitat preferences, not predation. Oecologia 68, 205–209.

Bergström, U., Englund, G., Bonsdorff, E., 2002. Small-scale spatial structure of Baltic Seazoobenthos-inferring processes from patterns. J. Exp. Mar. Biol. Ecol. 281, 123–136.

Bologna, P.A.X., Heck Jr., K.L., 2000. Impacts of seagrass habitat architecture on bivalvesettlement. Estuaries 23, 449–457.

Bologna, P.A.X., Heck, K.L., 2002. Impact of habitat edges on density and secondaryproduction of seagrass-associated fauna. Estuaries 25, 1033–1044.

Bonsdorff, E., Norkko, A., Boström, C., 1995. In: Eleftheriou, A., Ansell, A.D., Smith, C.J.(Eds.), Recruitment and population maintenance of the bivalve Macoma balthica(L.)-factors affecting settling success and early survival on shallow sandy bottoms.Proc. 28th Eur. Mar. Biol. Symp. Olsen & Olsen, Fredensborg, Denmark, pp. 253–260.

Boström, C., Bonsdorff, E., 2000. Zoobenthic community establishment and habitatcomplexity-the importance of seagrass shoot-density, morphology and physicaldisturbance for faunal recruitment. Mar. Ecol. Prog. Ser. 205, 123–138.

Boström, C., Bonsdorff, E., Kangas, P., Norkko, A., 2002. Long-term changes of a brackish-water eelgrass (Zostera marina L.) community indicate effects of coastaleutrophication. Est. Coast. Shelf Sci. 55, 795–804.

Boström, C., Jackson, E.L., Simenstad, C.A., 2006. Seagrass landscapes and their effect onassociated fauna: a review. Est. Coast. Shelf Sci. 68, 383–403.

Bowen, J.E., Hunt, H.L., 2009. Settlement and recruitment patterns of the soft-shell clam,Mya arenaria, on the northern shore of the Bay of Fundy. Can. Est. Coast. 32, 758–772.

Bremner, J., Frid, C.L.J., Rogers, S.I., 2003a. Assessing marine ecosystem health: the long-term effects of fishing on functional biodiversity in North Sea benthos. Aquat.Health & Manage. 6, 131–137.

Bremner, J., Rogers, S.I., Frid, C.L.J., 2003b. Methods for describing ecological functioningof marine benthic assemblages using biological traits analysis (BTA). Ecol. Indic. 6,609–622.

Bremner, J., Rogers, S.I., Frid, C.L.J., 2006. Methods for describing ecological functioningof marine benthic assemblages using biological trait analysis (BTA). Ecol. Indic. 6,609–622.

Butman, C.A., Grassle, J.P., Webb, C.M., 1988. Substrate choices made by marine larvaesettling in still water and in a flume flow. Nature 333, 771–773.

Charvet, S., Kosmala, A., Statzner, B., 1998. Biomonitoring through biological traits ofbenthic macroinvertebrates: perspectives for a general tool in streammanagement.Arch. Hydrobiol. 142, 415–532.

Chevenet, F., Soledec, S., Chessel, D., 1994. A fuzzy coding approach for the analysis oflong-term ecological data. Freshw. Biol. 43, 277–296.

Commito, J.A., Tita, G., 2002. Differential dispersal rates in an intertidal meiofaunacommunity. J. Exp. Mar. Biol. Ecol. 268, 237–256.

Commito, J.A., Celano, E.A., Como, H.J., Johnson, C.P., 2005. Mussels matter: postlarvaldispersal dynamics altered by a spatially complex ecosystem engineer. J. Exp. Mar.Biol. Ecol. 316, 133–147.

Connell, J.H., 1985. The consequences of variation in initial settlement vs. post-settlementmortality in rocky intertidal communities. J. Exp. Mar. Biol. Ecol. 93, 11–45.

Eggleston, D.B., Etherington, L.L., Elis, W.E., 1998. Organism response to habitatpatchiness: species and habitat-dependent recruitment of decapod crustaceans. J.Exp. Mar. Biol. Ecol. 223, 111–132.

Eggleston, D.B., Elis, E.W., Etherington, L.L., Dahlgren, C.P., Posey, M.H., 1999. Organismresponses to habitat fragmentation and diversity: habitat colonization by estuarinemacrofauna. J. Exp. Mar. Biol. Ecol. 236, 107–132.

Frid, C.L.J., Paramor, O.A.L., Brockington, S., Bremner, J., 2008. Incorporating ecologicalfunctioning into the designation and management of marine protected areas.Hydrobiologia 606, 69–79.

Gacia, E., Granata, T.C., Duarte, C.M., 1999. An approach to measurement of particle fluxand sediment retention within seagrass (Posidonia oceanica) meadows. Aquat. Bot.65, 255–268.

Gaines, S.D., Gaylord, B., Gerber, L.R., Hastings, A., Kinlan, B.P., 2007. Connecting places:the ecological consequences of dispersal in the sea. Oceanography 20, 90–99.

117C. Boström et al. / Journal of Experimental Marine Biology and Ecology 390 (2010) 106–117

Gambi, M.C., Nowell, A.R.M., Jumars, P.A., 1990. Flume observations on flow dynamics inZostera marina (eelgrass) beds. Mar. Ecol. Prog. Ser. 61, 159–169.

Hasegawa, N., Hori, M., Mukai, H., 2008. Seasonal changes in eelgrass functions: currentvelocity reduction, prevention of sediment resuspension, and control of sediment–watercolumn nutrient flux in relation to eelgrass dynamics. Hydrobiologia 596, 387–399.

Hewitt, J.E., Thrush, S.F., Dayton, P.D., Bonsdorff, E., 2007. The effect of scale on empiricalstudies of ecology. Am. Nat. 169, 398–408.

Hewitt, J.E., Thrush, S.F., Dayton, P.D., 2008. Habitat variation, species diversity andecological functioning in a marine system. J. Exp. Mar. Biol. Ecol. 366, 116–122.

Hily, C., Bouteille, M., 1999. Modifications of the specific diversity and feeding guilds inan intertidal sediment colonized by an eelgrass meadow (Zostera marina) (Brittany,France). C. R. Acad. Sci. Paris Sci. Vie Life Sci. 322, 1121–1131.

Hinchey, E.K., Nicholson, M.C., Zajac, R.N., Irlandi, E.A., 2008. Preface: marine and coastalapplications in landscape ecology. Land. Ecol. 23, 1–5.

Hovel, K.A., Fonseca, M.S., 2005. Influence of seagrass landscape structure on thejuvenile blue crab habitat–survival function. Mar. Ecol. Prog. Ser. 300, 179–191.

Hunt, H.L., Scheibling, R.E., 1997. Role of early post-settlement mortality in recruitmentof benthic marine invertebrates. Mar. Ecol. Prog. Ser. 155, 269–301.

Huynh, H., Feldt, L.S., 1976. Estimation of the box correction for degrees of freedomfrom sample data in the randomized block split plot designs. J. Educ. Stat. 1, 69–82.

Jennings, L.B., Hunt, H.L., 2009. Distances of dispersal of juvenile bivalves (Mya arenaria(Linnaeus), Mercenaria mercenaria (Linnaeus), Gemma gemma (Totten)). J. Exp.Mar. Biol. Ecol. 376, 76–84.

Johannesson, K., 1988. The paradox of Rockall: why is a brooding gastropod (Littorinasaxatilis) more widespread than on having a planktonic larval dispersal stage (L.Littorea)? Mar. Biol. 99, 507–513.

Johnson, G.E.L., Attrill, M.J., Sheehan, E.V., Somerfield, P.J., 2007. Recovery of meiofaunacommunities following mudflat disturbance by trampling associated with crab-tiling. Mar. Environ. Res. 64, 409–416.

Junkins, R., Kelaher, B., Levinton, J., 2006. Contributions of adult oligochaete emigrationand immigration in a dynamic soft-sediment community. J. Exp. Mar. Biol. Ecol.330, 208–220.

Kinlan, B.P., Gaines, S.D., 2003. Propagule dispersal in marine and terrestrialenvironments: a community perspective. Ecology 84, 2007–2020.

Magurran, A.E., 1991. Ecological diversity and its measurements. Chapman & Hall,London. pp. 91–99.

Martel, A., Chia, F.S., 1991. Drifting and dispersal of small bivalves and gastropods withdirect development. J. Exp. Mar. Biol. Ecol. 150, 131–147.

McGill, B.J., Enquist, B.J., Weiher, E., Westoby, M., 2006. Rebuilding community ecologyfrom functional traits. TREE. 21, 178–185.

Menge, B.A., Sutherland, J.P., 1987. Community regulation: variation in disturbance,competition, and predation in relation to environmental stress and recruitment.Am. Nat. 130, 730–757.

Munroe, D., McKinley, R.S., 2007. Effect of predator netting on recruitment and growthof manila clams (Venerupis philippinarum) on soft substrate intertidal plots inBritish Columbia. Can. J. Shellfish Res. 26, 1035–1044.

Orth, R.J., 1992. A perspective on plant–animal interactions in seagrasses: physical andbiological determinants influencing plant and animal abundance. In: John, D.M.,Hawkins, S.J., Price, J.H. (Eds.), Plant–animal interactions in marine benthos.Clarendon Press, Oxford, pp. 147–164.

Palmer, M.A., 1988. Dispersal of marine meiofauna: a review and conceptual modelexplaining passive transport and active emergence with implications forrecruitment. Mar. Ecol. Prog. Ser. 48, 81–91.

Peterson, C.H., 1986. Enhancement ofMercenaria mercenaria densities in seagrass beds:is pattern fixed during settlement season or altered by subsequent differentialsurvival? Limnol. Oceanogr. 31, 200–205.

Peterson, C.H., Luettich Jr., R.A., Micheli, F., Skilleter, G.A., 2004. Attenuation of waterflow inside seagrass canopies of differing structure. Mar. Ecol. Prog. Ser. 268, 81–92.

R Developmental Core Team, 2009. R: A Language and Environment for StatisticalComputing. R Foundation for Statistical Computing, Vienna, Austria3-900051-07-0.URL http://www. R-project.org.

Robbins, B.D., Bell, S.S., 1994. Seagrass landscapes: a terrestrial approach to the marinesubtidal environment. Trends Ecol. Evol. 9, 301–304.

Smith, C.R., Brumsickle, S.J., 1989. The effects of patch size and substrate isolation oncolonization modes and rates in an intertidal sediment. Limnol. Oceanogr. 34,1263–1277.

Sogard, S.M., 1989. Colonisation of artificial seagrass by fish and decapod crustaceans:importance of natural eelgrass. J. Exp. Mar. Biol. Ecol. 133, 15–37.

Tamaki, A., 1987. Comparison of resistivity to transport by wave action in severalpolychaete species on an intertidal sand flat. Mar. Ecol. Prog. Ser. 37, 181–189.

Tanner, J.E., 2003. Patch shape and orientation influences on seagrass epifauna aremediated by dispersal abilities. Oikos 100, 517–524.

Todd, C.D., 1998. Larval supply and recruitment of benthic invertebrates: do larvaealways disperse as much as we believe? Hydrobiologia 375 (376), 1–21.

Valanko, S., Norkko, A., Norkko, J., 2010. Strategies of post-larval dispersal in non-tidalsoft-sediment communities. J. Exp. Mar. Biol. Ecol. 384, 51–60.

von Ende, C.N., 1993. Repeated-measures analysis: growth and other time-dependentmeasures. Scheiner, S., Gurevitch, J.ŽEds., Design and Analysis of EcologicalExperiments. Chapman & Hall, London, In. pp. 113–137.

Warry, F.Y., Hindell, J.S., MacReadie, P.I., Jenkins, G.P., Connolly, R.M., 2009. Integratingedge effects into studies of habitat fragmentation: a test using meiofauna inseagrass. Oecologia 159, 883–892.

Waycott, M., Duarte, C.M., Carruthers, T.J.B., Orth, R.J., Dennison, W.C., Olyarnik, S.,Calladine, A., Fourqurean, J.W., Heck Jr., K.L., Hughes, R.A., Kendrick, G.A.,Kenworthy, W.J., Short, F.T., Williams, S.L., 2009. Accelerating loss of seagrassesacross the globe threatens coastal ecosystems. Proc. Nat. Acad. Sci. 106,12377–12381.

Whitlatch, R.B., Lohrer, A.M., Thrush, S.F., Pridmore, R.D., Hewitt, J.E., Cummings, V.J.,Zajac, R.N., 1998. Scale-dependent benthic recolonization dynamics: life stage-based dispersal and demographic consequences. Hydrobiologia 375 (376),217–226.

Whittaker, R.H., 1960. Vegetation of the Siskiyou Mountains, Oregon and California.Ecol. Monogr. 26, 1–80.

Zajac, R.N., 2008. Challenges in marine, soft-sediment benthoscape ecology. Land. Ecol.23, 7–18.

Zajac, R.N., Whitlatch, R.B., Thrush, S.F., 1998. Recolonization and succession in soft-sediment infaunal communities: the spatial scale of controlling factors. Hydro-biologia 375 (376), 227–240.