Vol. 3 8 3 :1 8 7 -1 9 7 ,2 0 0 9doi: 10.3354/meps07982

MARINE ECOLOGY PROGRESS SERIES Mar Ecol Prog Ser Published M ay 14

OPEN .ACCESS

Mechanisms of invasion resistance: competition among intertidal mussels promotes establishment of invasive species and displacement of native species

J. S. Shinen1*, S. G. Morgan2

'E stación Costera de Investigaciones M arinas, D epartam ento de E cología , Pontificia Universidad C atólica de Chile,A lam eda 340, Santiago, Chile

2B odega M arine Laboratory, Departm ent of Environm ental Scien ce and Policy, University of California Davis, PO Box 247,B odega Bay, C alifornia 94923, USA

ABSTRACT: U nderstanding interactions betw een invasive species and recipient communities is essential to determ ining w hether invasive species will becom e established and spread. In this study, we explored the role of competition and the specific m echanisms of interaction in limiting the spread of the M editerranean bay mussel M ytilus galloprovincialis w ithin a Pacific Northwest invasion front. We exam ined the role of direct (interference) and indirect (exploitation) m echanisms of competition am ong M. galloprovincialis and 2 native mussels (M. trossulus and M. californianus). As the fastest- growing organism s are often competitively dom inant in space-lim ited systems, such as rocky in ter­tidal communities, we used changes in relative perform ance (growth and survival) in monocultures and polycultures to assess interactions am ong mussels. Performance of M. galloprovincialis was always g reater than that of the 2 native species of mussels in both field and laboratory manipulations of species composition and density, indicating that interspecific competition did not strongly limit the grow th or survival of the invader. Moreover, the presence of M. galloprovincialis consistently led to both reduced grow th and survival of M. trossulus. Laboratory studies of mussel feeding and behavior revealed M. galloprovincialis to be a robust interference competitor. The invader restricted m ove­ment, sm othered and interfered with filter feeding of the 2 native species of mussels. Rather than lim­iting invasion, interference competition gave M. galloprovincialis a competitive advantage over the native mussels. Our results suggest M. galloprovincialis may have contributed to the displacem ent of M. trossulus along m uch of its historic southern range.

KEY WORDS: Exploitation competition • Interference competition • M ytilus • Rocky intertidalcommunities

---------------------------------- Resale or republication not permitted without written consent of the publisher---------------------------------

INTRODUCTION

A major ecological concern is the ability of com m uni­ties to resist invasion by an exotic species. The classical hypothesis states that as species diversity increases, communities becom e increasingly saturated, leaving few available niches that can be colonized by invasive species (Elton 1958). Empirical studies have found both positive (Naeem et al. 2000, Stachowicz et al. 2002) and negative (Levine & D'Antonio 1999, Stohlgren et al. 1999) relationships betw een diversity and invasibil-

ity, yet these conclusions appear to depend heavily on the system, species, and scale being investigated (Fri­dley et al. 2007). Discrepancies may be attributed to a lack of understanding of the m echanisms that generate local patterns of community structure w ithin invaded systems, such as interspecific competition, productiv­ity, or environm ental heterogeneity (Moore et al.2001). However, predicting invasion resistance depends on understanding how m echanisms m aintain­ing diversity suppress or facilitate an invader (Levine 2000). The ability of an exotic species to successfully

’Email: jlsh inen@ bio.puc.cl © In ter-R esearch 2009 • w w w .in t-res.com

188 M ar Ecol P rog Ser 383: 187-197, 2009

invade a new community may depend on the strength and form of its competitive interactions w ith species in the recipient community (Morrison 2000, Brown et al. 2002). Although m any studies of the interspecific com ­petition betw een native and invasive species have dem onstrated strong interspecific effects, com para­tively few studies identify or directly exam ine the m echanisms of interaction (but see Callaway & A schehoug 2000, Duyck et al. 2006), and exam ples outside plant ecology are rare.

In m arine systems, invasions of non-native species into resident communities are both contributions and threats to local biodiversity (Carlton 1996, Grosholz2002), but competitive interactions betw een marine native and invasive species are poorly understood (but see Castilla et al. 2004). The M editerranean bay m us­sel M ytilus galloprovincialis is a global invader with established populations in the Americas, Africa, A us­tralasia, and Jap an (McDonald et al. 1991, Sanjuan et al. 1997). The invasion of M. galloprovincialis has been particularly well docum ented in South Africa, w here the non-native m ussel has had w ide-ranging effects on entire communities along exposed, w ave-sw ept shores (reviewed by Branch & Steffani 2004). M ulti-trophic effects include competitive exclusion of native mussels (Van Erkom Schurink & Griffiths 1990) and herbivo­rous limpets (Steffani & Branch 2005), but the specific m echanisms of interaction betw een native mussels and M. galloprovincialis are not well understood.

Relatively little is known about the ecological influ­ence of M ytilus galloprovincialis on native com m uni­ties along the Pacific coast of North America. Since its introduction to San Diego harbor in the 1930s via ship­ping and aquaculture (Wonham 2004), M. galloprovin­cialis has spread and persisted throughout southern California (Geller 1999). A retrospective analysis of m useum collections revealed that the invader may have displaced the native mussel, M. trossulus, throughout m uch of the southern portion of its range (Geller 1999). The invasion front is currently located along the California coast from north of Point C oncep­tion to south of C ape M endocino. Within this region, M. galloprovincialis has become well established in protected bays and harbors but does not reach sub­stantial num bers in open coast rocky intertidal com m u­nities (Geller 1999, Rawson et al. 1999). The exposed coasts of the region also receive sympatric recruitm ent of M. galloprovincialis, and 2 native mussels, M. tros­sulus and M. californianus (reviewed in Braby & So­mero 2006). M. californianus numerically dom inates local rocky intertidal communities, and adult M. tros­sulus and M. galloprovincialis are rare and dispropor­tionately scarce relative to the num bers of settlers.

The disjunction betw een the larval recruitm ent p a t­terns and the adult mussel community structure within

the Pacific Northwest invasion front indicates that post-settlem ent mortality partially constrains the ongo­ing invasion of M ytilus galloprovincialis north of Point Conception (Johnson & Geller 2006). Predation, partic­ularly by the w helk Nucella ostrina, restricts the distri­bution of both M. trossulus and M. galloprovincialis to some extent, but predation rates do not appear to be sufficient to limit invasion (Shinen et al. 2009). M ore­over, a large settlem ent event of M. galloprovincialis could swamp the biotic resistance provided by p red a­tors (Hollebone & Hay 2007).

Competition may also play an im portant role in the invasion of M ytilus galloprovincialis. In rocky intertidal communities, w here available substrate for sessile or­ganism s can be limiting, the fastest-growing species is often the competitive dominant. Early work by H arger (1968) conducted on a possible mixture of all 3 species of M ytilus suggests that competition may play an im ­portant role in mussel grow th and mortality. Similarly in South Africa, the presence of M. galloprovincialis is frequently associated with a decline in the native m us­sel Perna perna (Bownes & M cQuaid 2006, Hanekom 2008). M. galloprovincialis, like other filter-feeding b i­valve molluscs, may com pete for resources (sestonic food or attachm ent substrate) directly by physical in ter­ference (crowding or pushing) or indirectly through exploitation (Frechette & Despland 1999). A lthough primarily sessile, mussels do move using a modified m uscular foot while releasing and attaching byssal threads and can engage in or disengage from com peti­tive interactions. M ovement and behavior can be ex ­trem ely im portant in interference competition, w here crowding and sm othering can occur, as well as in ex ­ploitation competition, w here individuals may move to better exploit resources. In this study, we exam ined competitive interactions among the 3 species of mussels to determ ine w hether the 2 native mussels can resist further invasion by M. galloprovincialis. Specifically, w e established the existence of a competitive hierarchy am ong these mussels, and identified and quantified the m echanism s of competitive interaction.

MATERIALS AND METHODS

General approach. All experim ents w ere conducted at the Bodega M arine Laboratory (BML) or in the sur­rounding Bodega M arine Reserve (BMR) in northern California, USA. The region is characterized by mixed sem idiurnal tides and strong oceanographic up- welling. All 3 species of mussels occur on exposed rocky shores of the BMR, although M ytilus californi­anus predom inates.

In all experim ents, changes in relative growth and survival rates am ong mussel m onocultures and poly­

S h in en & M organ: M echan ism s of com petition b e tw e e n n a tiv e an d invasive m ussels 189

cultures w ere used to estim ate competitive ability (Connell 1983). Since small settlers of M ytilus trossulus and M. galloprovincialis can be difficult to differenti­ate morphologically (Martel et al. 2000), larger juvenile mussels that are readily distinguishable (length 26.8 ± 3.8 mm) w ere used in all experim ents. Field studies of competitive interactions am ong more than 2 species are uncommon because they becom e unwieldy w hen all possible density controls for intraspecific com peti­tion are included. M. galloprovincialis and M. trossulus of appropriate size also w ere limiting, because thou­sands w ere needed for experim ents. These constraints compelled us to combine a design that is ideal for d is­tinguishing betw een interspecific and intraspecific competition w ith a replacem ent series design. Our substitutive design held total m ussel density constant while species composition varied (Table 1,A). T reat­m ents included 3 monocultures, 3 polycultures with 2 species (1:1), and 1 polyculture w ith 3 species (1:1:1). The substitutive use of 3 species am ong 2-species poly­cultures revealed species-specific effects betw een competitors while holding intraspecific effects con­stant. Our objective was to identify the possible exis­tence of a competitive hierarchy am ong the 3 species rather than the absolute differences betw een in terspe­cific and intraspecific effects, thereby reducing the need for intraspecific controls. However, 3 reduced- density monocultures (4 On) w ere included to estimate intraspecific effects in the field. M onoculture densities ideally would have been equal to the proportions within each of the polycultures (Underwood 1986), but mussels transplanted in low density at BMR often suf­fer high mortality m aking this intractable. M ussels are gregarious and need sufficient attachm ent matrix

within m ussel aggregations to prevent being swept from the substrate. The intraspecific m onoculture d en ­sity was selected to approxim ate the small mussel patches observed at BMR while ensuring transplant survival. A lthough the intraspecific control was not directly com parable to the polyculture treatm ents, p re ­cluding a determ ination of the absolute interaction strength, comparisons betw een high- and low-density m onocultures revealed the approxim ate direction and m agnitude of intraspecific effects.

To m eet statistical assumptions, grow th am ong spe­cies and within treatm ents was analyzed using the ANOVA of untransform ed and rank-transform ed data. Only probability values that w ere significant (p < 0.05) for both tests w ere accepted (Zar 1974). ANOVA was also perform ed on the arcsine square root of survivor­ship am ong species and w ithin treatm ents. Post-hoc analyses w ere m ade w ith both Tukey's HSD multiple comparisons (a = 0.05) (differences betw een all trea t­ments) and Dunnett's tests (a = 0.05) (using m onocul­ture groups as controls).

Competitive hierarchy of native and invasive mus­sels in the field. The competitive hierarchy of the 3 species of mussels was determ ined in the field beg in ­ning in May 2004. M ussels w ere collected by hand from naturally occurring m id-intertidal m ussel popula­tions. M ytilus californianus w ere collected from BMR (38° 18'18" N, 123° 3 '8" W). M. trossulus w ere collec­ted from Straw berry Hill, Oregon (44° 15' 385" N, 124° 07' 594" W), well north of the zone of sympatry of the 3 species of mussels. M. galloprovincialis w ere col­lected from Shell Beach in Tómales Bay, California (38° 06' 59.75" N, 122° 52' 23.93" W), w here the popula­tion is composed almost entirely of M. galloprovincialis

T able 1. E x p erim en tal d esig n for m o n o cu ltu res a n d p o lycu ltu res of M ytilu s californianus (C), M . trossulus (T), a n d M . ga llo ­provincia lis (G). T ab le show s n u m b ers of m usse ls in e ach trial. W ithin eac h study, colum ns re p re se n t trea tm en ts , excep t for m o n ocu ltu re trea tm en ts , w h e re e ach spec ies w as g ro w n separa tely . R ed u ced -d en sity m onocu ltu res w e re only in c lu d ed in th e

2004 field study. T he 2006 exp lo itation s tudy in c lu d ed a contro l trea tm e n t w h e re no m usse ls w e re p re se n t

E xperim en t M ussel sp. R educed- d ensity

m onocu ltu re

H igh-density

m onocu ltu re

2-spp. po lycu ltu re

(C XT)

2-spp. po lycu ltu re

(C X G)

2-spp. po lycu ltu re

(G XT)

3-spp.po lycu ltu re(C X T X G)

(A) F ield study (2004) Total m ussels 40 60 60 60 60 60n = 6 Aí. californianus 40 60 30 30 - 20No. trea tm e n ts = 10 Aí. trossulus 40 60 30 - 30 20(20 X 20 X 10 cm) Aí. galloprovincia lis 40 60 - 30 30 20

(B) In terfe ren ce study Total m ussels 12 12 12 12 12(2005) n = 5 Aí. californianus (not done) 12 6 6 - 4No. trea tm e n ts = 7 Aí. trossulus (not done) 12 6 - 6 4(42 X 26 X 17 cm) Aí. galloprovincialis (not done) 12 - 6 6 4

(C) E xploitation study Total m ussels 0 (control) 6 6 6 6 6(2006) n = 6 Aí. californianus - 6 3 3 - 2No. trea tm e n ts = 8 Aí. trossulus - 6 3 - 3 2(42 X 26 X 17 cm) Aí. galloprovincia lis - 6 - 3 3 2

190 M ar Ecol P rog Ser 383: 187-197, 2009

(Sarver & Foltz 1993, Suchanek et al. 1997). To further reduce the inadvertent inclusion of non-target species or hybrids, each m ussel was individually inspected for species-specific morphological features (McDonald et al. 1991). M ussels that did not resem ble the target population w ere elim inated from the experim ent. All mussels w ere transported to BML, cleaned of any epibionts, and acclim ated in species-specific, flow­through seaw ater aquaria for 1 mo.

In June, the length, height, and w idth (mm) of m us­sels w ere recorded, and a small notch was m ade at the growing edge of each shell as a reference scar for new shell grow th (Coe & Fox 1942). Mussels w ere haphaz­ardly assigned to experim ental treatm ents, w hich w ere spatially random ized and outplanted to a previously cleared area of mussel bed in the middle intertidal m us­sel zone at the BMR. Stainless steel w ire m esh (7 mm) enclosures (20 x 20 x 10 cm) w ere constructed around each of the experim ental cultures and w ere bolted d i­rectly onto the rock. M ussels w ere held close to the rock surface for 2 w k by flexible plastic m esh (7 mm) to facilitate reattachm ent to the substrate by byssal threads. In August, the enclosures w ere removed, and all of the mussels and mussel shells w ere transported back to the BML. Growth and survivorship of each spe­cies of mussel in each treatm ent w ere calculated. Cages w ere needed to elim inate em igration of mussels and exclude high densities of predatory w helks and sea- stars. The predator-exclusion cages w ere not completely invulnerable to small whelks, but shells that showed signs of predation (i.e. w helk drill holes) w ere minimal (<5%), did not affect the relative densities of mussels, and w ere excluded from mortality estimates. Partial- cage controls w ere not conducted, because full p reven­tion of predation and emigration was essential and only relative changes in growth and survival in polyculture and monoculture w ere needed to interpret species interactions. Cages did not trap sediment or greatly re ­strict w ater flow on this highly energetic, w ave-swept shore and had little effect on mussel performance. In­deed, growth and survival of mussels w ere generally consistent in field and in laboratory experiments indicat­ing that competitive ability and species interactions w ere com parable in different flow regimes.

Interference competition in the laboratory. We also conducted an experim ent in the laboratory to d e te r­mine w hether physical interference is an im portant m echanism of competition am ong the 3 species of m us­sels. A lthough laboratory conditions differ from those in the field, mussel behaviors during immersion are impossible to monitor in field conditions. In Decem ber 2004, mussels of each species w ere haphazardly assigned to m onoculture and polyculture treatm ents (Table 1,B). The total num ber of mussels in each trea t­m ent w as reduced to 12 to simplify the observation of

mussel behaviors while m aintaining the same relative proportions of species in the cultures that w ere used in the field. The reduced-density m onoculture treatm ents w ere eliminated. Replacem ent series-type designs such as this can be valuable comparative tools (Cou- sens & O'Neill 1993), w here consistent patterns of m us­sel grow th and survivorship betw een laboratory and field treatm ents may indicate conserved, density- independent competitive interactions betw een species (Jolliffe 2000). Each mussel w as notched, and length, height, and w idth w ere recorded before placing m us­sels into plastic containers (42 x 26 x 17 cm). M ussels in each container w ere arranged in a clump directly under a steady flow of sand-filtered seawater.

The experim ent was term inated after 5 wk. Photo­graphs w ere taken, and individual m ussel position was noted in each of the m onoculture and polyculture source containers. M ortality was recorded and growth increm ents above the filed notch w ere m easured. The distance traveled by mussels from the original p lace­m ent of clum ped mussels was calculated w ith Im ageJ software (Abramoff et al. 2004). Data w ere analyzed w ith ANOVA and post-hoc analyses w ere perform ed w ith both Tukey's HSD multiple comparisons (a = 0.05) (differences betw een all treatm ents) and D unnett's tests (a = 0.05) (using m onoculture groups as controls). Species-specific daily grow th rates w ere also pooled over all treatm ents regressed against the total distance traveled (cm) from the central mussel clump.

Exploitation competition in the laboratory. A series of feeding trials was conducted in the laboratory to determ ine w hether the 3 species of mussels compete for food. As a proxy for filter-feeding ability, clearance rates of mussel cultures w ere assessed using 2 m eth­ods. First, we m onitored the relative decline in concen­tration of in vivo chlorophyll a (chi a) over time. Work on related mussels has dem onstrated that clearance rates are similar in both laboratory and in situ condi­tions (Velasco & Navarro 2005). In April 2006, small mussels (2 to 4 cm length) of the 3 species of mussels w ere collected from the same collection sites over a single spring tide series and w ere transported to BMR. M ussels w ere acclim ated in flow-through seaw ater for 1 mo before being assigned haphazardly to m onocul­ture and polyculture treatm ents (Table 1,C). To limit movement, mussels w ere placed in small (10 cm diam ­eter) glass culture dishes, which w ere subm erged in the center of larger plastic containers (42 x 26 x 17 cm) containing 11 liters of static, sand-filtered seawater. Each plastic container was supplied w ith a single w eighted airline and air stone. Cultures w ere posi­tioned in a random ized block design inside large tubs (110x110x55 cm). A 10 cm tail standpipe and running seaw ater in the large tubs kept the smaller plastic con­tainers at am bient seaw ater tem perature.

S h in en & M organ: M echan ism s of com petition b e tw e e n n a tiv e an d invasive m ussels 191

25 -,

20 -

15 -

10 -

5 -

0

Feeding trials lasted 6 h, approxim ating the average immersion time for mussels in a sem idiurnal tidal cycle. Concentrations of chi a w ere m easured with a handheld fluorometer (Turner Designs Aquafluor). No-mussel controls w ere included in each replicate tub to monitor natural fluctuations in chi a during feeding trials. Initial w ater samples w ere taken before mussel cultures w ere added to containers and at 60 min intervals. M ean chi a concentrations per in ter­val w ere quantified for each container by taking w ater samples (1 ml) at 3 locations (top, middle, and bottom), at m id-depth in the aquaria. At the term ina­tion of the feeding trials, m ussel tissue was dissected and dried at 60°C for a minimum of 24 h at the end of the experim ent to determ ine the total dry mass of mussels in each culture.

C learance rates of mussels in m onocultures and polycultures w ere also assessed in term s of the total particulate m atter (TPM) and the partic­ulate organic m atter (POM) rem oved over the course of the feeing trials. Changes in concentra­tion of TPM and POM w ere estim ated by taking 2 replicate w ater samples (1 1) before and after feeding trials. Samples w ere vacuum filtered onto pre-ashed, tared W hatm an GF/C filters.TPM was determ ined after 24 h of drying at 60°C.Total organic m atter was calculated from the dif­ference betw een the TPM and the filter weight after samples w ere ashed at 400°C for 4 h.

Analyses w ere perform ed on the total p er­centage change in concentration of chi a, TPM, and POM at the end of each feeding trial. Values w ere standardized by the total dry tissue mass of mussels (g) in each culture. No blocking effects w ere found; thus, differences in filter feeding b e ­tw een cultures w ere com pared using ANOVA, and post-hoc analyses w ere performed with both Tukey's HSD multiple comparisons (a = 0.05) (differences betw een all treatm ents) and Dun- nett's tests (a = 0.05) (using monoculture groups as controls).

RESULTS

Competitive hierarchy of native and invasive mussels

Growth and survival in field monocultures and polycultures suggest that M ytilus galloprovin­cialis is the dom inant competitor in a space-lim ­ited competitive hierarchy (Fig. 1). The invader M. galloprovincialis grew 10 times faster than ei­ther of the 2 native species (F2i2371 = 688.34, p < 0.001). Additionally, the presence of M. gallo-

o0)CÖ

-t—'

CLCÖO0

CL

provincialism a 2-species polyculture with M. trossulus reduced grow th and survival of M. trossulus, indicating that the invader can competitively displace this native species. M. californianus was weakly positively affected by interspecific interactions.

In addition to establishing the competitive robustness of M ytilus galloprovincialis, other interspecific and in ­traspecific effects w ere observed in cultures. However, these effects were secondary, since the strengths of these interactions w ere generally weak. M. californianus grew most slowly and seem ed least affected by experim ental conditions and treatments. However, this species tended to benefit from decreasing density of conspecifics in the field. M. trossulus was influenced by competition b e ­tw een conspecifics and the native competitor, M. califor­nianus. The invader, M. galloprovincialis, w as influ-

M. californianus

ñ n n

120 -i

100 -

80 -

60 -

40 -

20 -

0 -

M. californianus

E 253 .0 204—»0

15

10

0

CR C CT CG CTG

M. trossulus 1 2 0 -i

1 0 0 -

> 80 -

¿ 6 0 ̂

Sn 40 -

n r " i20 -

0

CR C CT CG CTG

M. trossulus

TR T CT TG CTG

M. galloprovincialis

TR T CT TG CTG

M. galloprovincialis

Fig. 1. C om petitive h ierarchy experim ent. M ean pe r capita g row th ra tes an d p e rcen tag e survival of 3 species of m ussels w ith in m onocultures and polycultures: C = M ytilus californianus, T = Aí. trossulus, G = Aí. gallo­provincialis. R = red u ced density. Bars ind icate +1 SE. 'G roups th a t differed significantly (p < 0.05) from th e h igh-density m onoculture controls b a sed on D unnett's post-hoc analyses of bo th th e un transfo rm ed and

rank-transform ed da ta

192 M ar Ecol P rog Ser 383: 187-197, 2009

enced by changes in total mussel density rather than competition betw een hetero- or con­specifics. The growth rate of the invader decrea­sed only w here total mussel density was re ­duced: in reduced-density m onoculture and in field polyculture w ith M. trossulus, w here M. trossulus experienced high mortality. Lastly, all 3 species experienced m oderately higher p er­formance in the 3-species polyculture than in 2 -species polycultures, suggesting the mussels may benefit weakly from an indirect facilitation, particularly diffuse competition. Several of these tendencies w ere non-significant (p > 0.05), and none alter our m ain finding that the invader is a strong interference competitor.

Interference competition

75 -,

60 -

45 -

30 -

15 -

M. californianus

(0)

CT CG CTG■öE3-CD05

75 i

60 -

45 -oO)CÖ

-t—'

CLCÖO

30 -

oCL

15 -

75 n

Relative grow th rates of the different mussels in the laboratory w ere consistent w ith those observed in the field (Fig. 2). Invasive M ytilus galloprovincialis grew the most, followed by M. trossulus and then M. californianus (F2i408 =138.78, p < 0.001). The grow th rate of native M. trossulus decreased in polyculture with invasive M. galloprovincialis, as it did in the field. Survival was generally high across all species and treatm ents (Fig. 2); however, all of the observed mortality occurred w ithin mussel clumps w here individuals appeared to be sm othered by other mussels that had climbed atop the mussel clump. M ortality of native M. californianus was greatest overall (-15% total, F2i57 = 5.70, p = 0.006).

All 3 species of mussels moved inside the plas­tic aquaria, and dispersal frequency and distance in the treatm ents varied by species and trea t­m ent (Fig. 3). Overall, M ytilus trossulus em i­grated from mussel clumps more frequently (F2,57 = 5.79, p = 0.005) and farther (F2 4 0 8 = 10.44, p < 0.001) than the other 2 species. The growth rate of M. trossulus had a m oderately positive relationship to the distance traveled from the mussel clump (R2 = 0.24, p < 0.001). However, only w eakly positive relationships w ere observed betw een grow th rate and distance trav­eled by native M. californianus (R2 = 0.12, p < 0.001) and invasive M. galloprovincialis (R2 = 0.02, p < 0.001).

Exploitation competition

Differences in clearance rate w ere observed among species and treatm ents in feeding trials (Fig. 4). Surpris­ingly, the m ussel w ith the greatest grow th rate, M ytilus

M. trossulus

CT TG CTG

M. galloprovincialis

120-,

100-

80 -

60 -

40 -

20 -

0 -

120-,

100-m> 80->-¡co 60-0s*

40-

20-

0 -

120-,

100-

80 -

60 -

40 -

20 -

0 -

M. californianus

CT CG CTG

M. trossulus

CT TG CTG

M. galloprovincialis

Fig. 2. In te rfe ren ce com petition experim en t. M ean p e r cap ita g row th ra te s a n d p e rc e n ta g e survival of 3 spec ies of m ussels w ith in m onocu l­tu res a n d po lycu ltu res. Bars in d ica te +1 SE. 'G ro u p s th a t d iffered sig ­n ificantly (p < 0.05) from th e m o n ocu ltu re controls b a se d on D u n n e tt's post-hoc analyses of b o th th e u n tran sfo rm ed a n d ran k -tran sfo rm ed data.

See Fig. 1 for spec ies abbrev ia tions

galloprovincialis, was least proficient. Of the m onocul­tures, only M. californianus and M. trossulus rem oved detectable amounts of chi a from the seston (F320 = 4.16, p = 0.019). Among polycultures, M. californianus com ­bined w ith M. trossulus and the 3-species polyculture rem oved the greatest chi a (F425 = 22.9, p < 0.001). No filter feeding was evident in 2 -species polycultures that included M. galloprovincialis. No differences in the percentage change of TPM am ong mussel cultures and the no-m ussel control w ere detected (F740 = 1.96, p = 0.085; Fig. 5). However, monocultures of M. califor­nianus and polycultures of M. californianus and M. trossulus rem oved the most POM from the seston (F7i40 = 4.25, p = 0.001).

S h in en & M organ: M echan ism s of com petition b e tw e e n n a tiv e a n d in v asiv e m usse ls 193

"ÖCD(/)0CL<0T3£=o

-4—'

oCLO

1.0

0.8

0.6

0.4

0.2

0.0

1 .0 i

0.8

0.6 -

0.4

0.2 -

0.0

M. californianus 10

6 -

4 -

2 -

(0)

M. californianus

_I_

CT CG CTG

M. trossulus 10 n

o o £=CÖ4—'u>~uT e 4 -<i.)oCL cob

6 -

2 -

C CT

M. trossulus

CG CTG

1.0

0.8

0.6

T CT TG

M. galloprovincialis

CTG10 n

6 -

T CT TG

M. galloprovincialis

CTG

:lG CG TG CTG

Fig. 3. D ispersal behav io rs am o n g n a tiv e a n d invasive spec ies of m ussels show ing th e p ro p o rtio n of m usse ls fo u n d ou tside of m usse l clum ps a n d th e m ea n d istan ce trav e le d by m usse ls from clum ps w ith in m onocu ltu res an d po lycu ltu res. Bars in d ica te +1 SE. 'G ro u p s th a t d iffered sign ifican tly (p < 0.05) from th e m o n ocu ltu re controls b a se d on D u n n e tt's post-hoc analyses.

See Fig. 1 for spec ies ab brev ia tions

DISCUSSION

Competitive hierarchy among mussels

Overall, we found compelling evidence that M yti­lus galloprovincialis is competitively dom inant over M. trossulus, w here the invader consistently reduced the grow th and survival of M. trossulus in both labora­tory and field polycultures. Although the aim of this study was to identify specific m echanisms of in terac­tion betw een mussels over relatively short time scales (4 to 8 wk), we w ere also able to detect population- level effects of competition with the invader as evi­denced by the high mortality of M. trossulus. The d u ra ­tion of our experim ental trials may have been too short

to determ ine w hether M. galloprovincialis is also competitively dom inant over M. californi­anus. However, the rapid grow th of the invader com bined with increased recruitm ent of M. galloprovincialis might easily over­whelm M. californianus in rocky intertidal communities w here the fastest-grow ing orga­nisms often dom inate the available substrate. Experim ental m ussel cultures in the labora­tory and field differed in fluid dynamics, immersion time, and total mussel density, but despite differences in experim ental condi­tions, M. galloprovincialis always ou tper­formed its native congeners.

Mechanisms of interspecific competition

The invader interacted with native mussels primarily through interference competition. M ytilus galloprovincialis sm othered hetero­specific competitors, limited their mobility, and im peded filter feeding. Rather than d en ­sity-m ediated overgrow th of heterospecifics, the tendency to aggregate or disperse ap ­peared to m ediate competition betw een native and invasive mussels. Even without recruitm ent and grow th of new individuals, mussels w ere able to smother one another by positioning them selves atop others. In the lab ­oratory, mussels that did not disperse from the m ussel clump incurred reduced grow th rates or higher mortality w hen cultured with M. galloprovincialis. While M. galloprovin­cialis and M. californianus tended to be re la ­tively non-mobile, M. trossulus dispersed from mussel clumps more frequently and at g reater distances. Performance was consis­tently higher am ong M. trossulus that w ere

found outside mussel clumps. However, M. trossulus dispersed less frequently in polyculture w ith M. gallo­provincialis, a likely result of M. trossulus being prevented from dispersing by the byssal threads of M. galloprovincialis.

Similar patterns of clumping, smothering, and dis­persal w ere observed in field m onocultures and poly­cultures. M ytilus trossulus w ere loosely assem bled with conspecifics and non-conspecifics, w hereas M. californianus and M. galloprovincialis typically w ere tightly aggregated (Fig. 6). H arger (1968) hypothesized that highly mobile behavior am ong cer­tain mussels may have had a definite benefit in calm er waters, w here crawling above competitors or substrate may be necessary to prevent suffocation or

194 M ar Ecol P rog Ser 383: 187-197, 2009

20Monocultures

Control

- 2 0 -

-40-0)= L -60->,szCLo1—

-80-

o -100-szo£= -120

0 2 3 4 61 500)£=CÖ

20 rPolycultures Control

CGCTGTCTG

szO

-40-

-60-

-80-

- 100 -

-1200 2 3 4 61 5

Elapsed tim e (h)

Fig. 4. E xploitation com petition experim ent. C hi a dep le tio n by 3 spec ies of m ussels. D a ta w ith ±SE b a rs a re th e av erag e p e rc e n ta g e ch an g e of in vivo chi a co ncen tra tion s ta n d a rd ­ized by to ta l m usse l d ry m ass. No m usse ls w e re p re se n t in

contro l trea tm en ts . See Fig. 1 for spec ies ab b rev ia tio n s

to find suitable areas for filter feeding. In contrast, dis- lodgem ent by wave force can be an im portant source of mortality in mussels, which is m ediated by ag g re­gation and increased attachm ent tenacity (Hunt & Scheibling 2002). The tendency to aggregate in M. galloprovincialis could reflect w eaker attachm ent ability or higher hydrodynamic loading of individuals (Rius & M cQuaid 2006). It may also explain the intraspecific density effects experienced by the invader in field cultures and the tendency of M. gallo­provincialis to attach quickly to neighboring mussels (pers. obs.). Additional physical forces may m ediate mussel movem ent and interference competition. In South Africa, competitive dynamics betw een the native Perna perna and invasive M. galloprovincialis vary in response to sand accum ulation (Zardi et al. 2008), upwelling intensity (Xavier et al. 2007), and wave forces (Rius & M cQuaid 2006).

In addition to physical interference, we found that mussel behaviors w ere also involved in exploitation of shared food resources. Indeed, we found further evi­dence of interference competition am ong the 3 mussel species indicating that M ytilus galloprovincialis was

0

- 2 0 -

-40-

TU) -60-c o■5 -8 0 -cCD

c - 1 0 0 -O B C T G CT CG TG CTG£ 0 -

0 0)

« - 2 0 - sz O

^ -40-

-60-

-80-

- 100 -

Fig. 5. F ilter fe ed in g on to ta l p a r ticu la te m atte r (TPM) an d p a r ticu la te o rgan ic m atte r (POM) by 3 spec ies of m ussels. D a ta a re th e av era g e p e rc e n ta g e ch an g e of concen tra tions s ta n d a rd ize d b y to ta l m usse l d ry m ass. B = control. E rror b a rs in d ica te +1 SE. 'G ro u p s th a t d iffered sign ifican tly (p < 0.05) from th e n o -m usse l controls b a se d on D u n n e tt's post-hoc

analyses. See Fig. 1 for spec ies ab b rev iatio n s

behaviorally m ediating chi a clearance rates in poly­cultures. The presence of M. galloprovincialis seem ed to stop the consumption of chi a altogether rather than simply reducing the clearance rate relative to its d en ­sity in polyculture. A lthough the scope of the present study was limited to revealing interference effects of M. galloprovincialis on native competitors, the seston clearance trials dem onstrated different feeding rates am ong the 3 species, suggesting that exploitation might be an im portant m echanism of competition. However, the mussel w ith the fastest grow th rate, M. galloprovincialis, exhibited the slowest feeding rate. These results w ere surprising and contradict other studies that have dem onstrated a strong ex ­ploitative advantage of M. galloprovincialis over other mussel competitors (Griffiths et al. 1992, Hilbish et al. 1994). M. galloprovincialis in the experim ents p re ­sented here may have been more occupied w ith byssal attachm ent than feeding, and clearance rates may increase w ith longer experim ental duration (Byrnes 2008).

B C T G CT CG TG CTG

S h in en & M organ: M echan ism s of com petition b e tw e e n n a tiv e an d invasive m ussels 195

Field Laboratory have dram atic community-level consequen­ces for rocky intertidal communities. In partic­ular, an extensive invasion of M. galloprovin­cialis may reduce the diversity and richness of epiphytic algae and sessile invertebrates asso­ciated w ith the shells of mussels (Shinen & M organ unpubl.). Furtherm ore, the differ­ences in filter-feeding rate betw een native and invasive mussels suggest that displace­m ent of M. trossulus by M. galloprovincialis may decrease or disrupt community biofiltra­tion rates, w hich may affect a variety of other native intertidal species. Invasive bivalves that alter phytoplankton distribution and abundance can have dram atic and u n ­predictable effects on multiple trophic levels and ecosystem scales (Noonburg et al. 2003, Kimmerer 2006). These results illustrate the im portance of elucidating interaction m echa­nisms to understand invasion resistance and anticipate the potential impacts of invasive species.

Despite our evidence that M ytilus gallo­provincialis is a strong competitor among native mussels, the invader is still only weakly established in central and northern California. The current extent of the invasion likely reflects a balance betw een predation and recruitm ent rates. Persistent oceanic

Fig. 6. Photographs ot rep resen ta tive m ussel clum ping behavior in field an d laboratory m onocultures ot (a) M ytilus californianus, (b) M. trossulus,

a n d (c) M. galloprovincialis

Consequences of interspecific competition

Interspecific competition has long been recognized as an im portant force regulating rocky intertidal com ­munity structure (Connell 1961), but it does not appear to be limiting the spread of M ytilus galloprovincialis. The rapid grow th rate, m echanical interference, and behavioral dom inance of M. galloprovincialis will likely enable further invasion in the Pacific Northwest. M. galloprovincialis appears to be contributing to the displacem ent of the native M. trossulus (Geller 1999). Regional extinctions caused by competition betw een native and invasive species are rare (Davis 2003), but increasing populations of M. galloprovincialis may

upwelling and offshore larval transport in the zone of mussel sympatry limits recruitm ent (Connolly et al. 2001) and may be limiting the establishm ent of M. galloprovincialis. On the west coast of South Africa, w here M. gallo­provincialis has been so profoundly success­ful, the high reproductive output of M. gallo­provincialis, coupled w ith larval retention, translates directly into dram atically high rates of recruitm ent to the shore (Harris et al. 1998). Additionally, predation is strong force limiting the invasion of M. galloprovincialis. The dog- w helk Nucella ostrina selectively eats M. gal­

loprovincialis, irrespective of the density of the invader relative to native mussels (Shinen et al. 2009). How­ever, aquaculture of M. galloprovincialis has increased exponentially since the 1950s (Food and Agriculture Organization of the United Nations 2007), increasing larval supply and recruitm ent potential far north of the current invasion front. Increases in the populations M. galloprovincialis will likely coincide w ith increases in the im portance of interspecific competition in resident rocky intertidal communities.

Our findings reinforce an em erging consensus am ong ecologists that invasion resistance cannot be predicted from simple rules and is instead a function of

196 M ar Ecol P rog Ser 383: 187-197, 2009

num erous interacting m echanisms (D'Antonio 1993, Thebaud et al. 1996). In the present study, closely related congeners all exhibited varying responses to inter- and intraspecific density effects that would not likely have been predicted without m anipulative experim ents that helped to identify the m echanism s of competition. These results may provide a cautionary tale for other invasion studies involving closely related natives and invaders, since relatedness is not always an accurate predictor of competitive ability and inva­siveness (Lambrinos 2002). Moreover, studies that use a comparative approach betw een closely related native and invasive species can provide valuable insights linking coexistence m echanisms and invasion success. For instance, understanding the m echanisms of interspecific interactions betw een native and inva­sive species may help to predict the likely impact of an invasion in a native community. Finally, discerning the m echanisms of competition betw een native and inva­sive species may also help direct m anagem ent efforts to limit the spread or reduce the impacts of exotic invaders.

A ck n o w led g e m e n ts . We th a n k M. A. H olyoak an d J. J. Sta- chow icz for com m enting on th e ex p erim en ta l d e sig n a n d an earlie r draft of th e m anuscrip t. Special th an k s to M. C. Vas- quez, B. G. M iner, J. L. F isher, A. J. M ace, E. A. Fairbairn , Y. H en n eb erry , J. E. B yrnes, a n d M. E. B rack en for th e ir in v a lu ab le adv ice a n d assistance in th e lab o ra to ry a n d th e field. T h an k s also to E. Sanford , J. Sones, a n d S. A. T hom pson for m u sse l collections.

LITERATURE CITED

A bram off MD, M ag e lh aes PJ, R am SJ (2004) Im age p ro cess­in g w ith Im ageJ. B iophotonics Int 11:36-42

B ow nes SJ, M cQ uaid CD (2006) Will th e invasive m usse l M ytilu s galloprovincia lis L am arck rep lace th e ind ig en o u s Perna p ern a L. on th e so u th coast of South Africa? J Exp M ar Biol Ecol 338:140-151

B raby CE, Som ero G N (2006) Ecological g rad ien ts an d re la ­tive ab u n d an c e of n a tiv e (M ytilus trossulus) a n d invasive (M ytilus galloprovincialis) b lu e m ussels in th e C alifornia h y b rid zone. M ar Biol 148:1249-1262

B ranch GM , Steffani C N (2004) C an w e p red ic t th e effects of a lien species? A case-h isto ry of th e invasion of South A frica b y M ytilu s galloprovincialis (Lam arck). J Exp M ar Biol Ecol 300:189-215

B row n BJ, M itchell RJ, G rah am SA (2002) C om petition for po llina tion b e tw e e n an invasive spec ies (purple loose­strife) an d a n a tiv e congener. Ecology 83:2328-2336

B yrnes JE (2008) P a tte rn s a n d con seq u en ces of ch an g in g b io ­diversity in coasta l m arin e food w ebs. PhD thesis, U niver­sity of C alifo rn ia Davis

C allaw ay RM, A sch eh o u g ET (2000) Invasive p lan ts versus th e ir n e w an d old neighbors: a m ech an ism for exotic in v a ­sion. Science 290:521-523

C arlton JT (1996) P a tte rn , p rocess, a n d p red ic tio n in m arin e invasion ecology. Biol C onserv 78:97-106

C astilla JC , G uiñez R, C aro AU, O rtiz V (2004) Invasion of a rocky in te rtid a l sho re by th e tu n ica te Pyura praepu tia lis in

th e Bay of A n to fagasta , C hile. Proc N atl A cad Sei USA 101:8517-8524

C oe WR, Fox DL (1942) Biology of th e C alifornia se a m usse l (M ytilus californianus). I. In fluence of tem p e ra tu re , food supply, sex, a n d ag e on th e ra te of grow th. J Exp Zool 90 :1 -30

C onnell JH (1961) E ffects of com petition an d p re d a tio n by Thais lapillus, an d o ther factors on n a tu ra l p o p u latio n s of th e b a rn ac le B alanus balanoides. Ecol M onogr 31:61-104

C onnell JH (1983) O n th e p re v a len c e a n d re la tiv e im p o rtan ce of in terspecific com petition: ev id en ce from field e x p eri­m en ts. A m N at 122:661-696

C onnolly SR, M en g e BA, R o u g h g a rd en J (2001) A la titu d in a l g rad ien t in rec ru itm en t of in te rtid a l in v erteb ra tes in th e n o rth ea s t Pacific O cean . Ecology 82:1799-1813

C ousens R, O 'N eill M (1993) D ensity d e p en d e n ce of re p la c e ­m en t se ries experim en ts. O ikos 66:347-352

D 'A ntonio CM (1993) M echan ism s contro lling in v asio n of coasta l p lan t com m unities by th e a lien su ccu len t C arpo­brotus edulis. Ecology 74:83-95

D avis M A (2003) Biotic globalization: does com petition from in tro d u c ed spec ies th re a te n biodiversity? B ioscience 53: 4 8 1 -4 8 9

D uyck PF, D avid P, Ju n o d G, B runei C, D upont R, Quilici S(2006) Im portance of com petition m echan ism s in su cces­sive invasions by po ly p h ag o u s tep h ritid s in La Reunion. Ecology 87:1770-1780

Elton CS (1958) T he ecology of invasions by anim als an d p lan ts. M eth u en , L ondon

Food a n d A g ricu ltu re O rg an iza tio n of th e U n ited N ations(2007) C u ltu red A quatic Species In form ation Program m e: M ytilu s galloprovincia lis (Lam ark, 1819). w w w .fao .o rg / f ish e ry /cu ltu red sp ecies /M y tilu s_ g a llo p ro v in cia lis/en

F rech ette M, D esp lan d E (1999) Im p aired shell g a p in g an d food d ep le tio n as m echan ism s of asym m etric com petition in m ussels. E coscience 6 :1-11

F rid ley JD , S tachow icz JJ , N a ee m S, Sax DF a n d o thers (2007) T he invasion paradox : reconciling p a tte rn a n d p rocess in spec ies invasions. Ecology 88 :3-17

G eller JB (1999) D ecline of a n a tiv e m usse l m ask ed by sib ling spec ies invasion. C onserv Biol 13:661-664

Griffiths CL, H ockey PAR, V an E rkom S churink C, Roux PJL (1992) M arine invasive a liens on South A frican shores: im plications for com m unity s tru c tu re a n d troph ic fu n c tio n ­ing . S Afr J M ar Sei 12:713-722

G rosholz E (2002) Ecological an d evo lu tionary con seq u en ces of coasta l invasions. T rends Ecol Evol 17:22-27

H an ek o m N (2008) Invasion of an in d ig en o u s Perna pern a m usse l b e d on th e so u th coast of South A frica b y a n a lien m usse l M ytilu s galloprovincia lis an d its effect on th e asso ­c ia ted farm a. Biol Invasions 10:233-244

H a rg er JRE (1968) T he ro le of b eh av io ra l traits in in fluencing th e d istribu tion of tw o spec ies of sea m ussel, M ytilu s edulis a n d M ytilu s californianus. V eliger 11:45-49

H arris JM , B ranch GM , Elliott BL, C urrie B a n d o thers (1998) S patia l a n d tem p o ra l variab ility in rec ru itm en t of in te r­tid a l m usse ls a ro u n d th e coast of so u th e rn Africa. S Afr J Zool 33:1-11

H ilb ish TJ, B ayne BL, D ay A (1994) G enetics of physio logical d ifferen tia tio n w ith in th e m arin e m usse l g en u s M ytilus. Evolution 48:267-286

H ollebone AL, H ay ME (2007) P ro p ag u le p re ssu re of an in v a ­sive crab overw helm s n a tiv e bio tic resis tan ce . M ar Ecol P rog Ser 342:191-196

H u n t HL, Scheib ling RE (2002) M ovem ent a n d w av e dis- lo d g em en t of m usse ls on a w av e-ex p o sed rocky shore. V eliger 45:273-277

S h in en & M organ: M echan ism s of com petition b e tw e e n n a tiv e an d invasive m ussels 197

Jo h n so n SB, G eller JB (2006) L arval se ttlem en t can exp la in th e ad u lt d istribu tion of M ytilu s californianus C o n rad b u t no t of Aí, galloprovincia lis L am arck or Aí, trossulus G ould in M oss L anding , c en tra l C alifornia: ev idence from genetic iden tification of spat. J Exp M ar Biol Ecol 328: 136-145

Jolliffe PA (2000) T he re p la ce m en t series. J Ecol 88:371-385 ^ . K im m erer W J (2006) R esponse of anchovies d am p en s effects

of th e invasive b ivalve Corbula am urensis on th e San Francisco E stuary foodw eb. M ar Ecol P rog Ser 324: 207-218

L am brinos JG (2002) T he v a riab le invasive success of Cor- taderia species in a com plex lan d scap e . Ecology 83: 5 1 8 -5 2 9

L evine JM (2000) Species d iversity a n d b io log ical invasions: re la tin g local p rocess to com m unity p a tte rn . Science 288: 852-854

^ . L evine JM , D 'A ntonio CM (1999) E lton rev isited : a rev iew of ev id en ce lin k in g d iversity a n d invasibility. O ikos 87 :15-26

^ . M arte l AL, A uffrey LM, R obles CD, H o n d a BM (2000) Id en ti­fication of se ttlin g a n d early p o stlarv a l s tag es of m ussels (M ytilus spp.) from th e Pacific coast of N o rth A m erica, u s in g p ro d issoconch m orphology an d genom ic DNA. M ar Biol 137:811-818

M cD onald JH , S eed R, K oehn RK (1991) A llozym es an d m orphom etric ch aracte rs of th re e spec ies of M ytilu s in th e n o rth e rn a n d so u th e rn hem isp h eres . M ar Biol 111: 323 -3 3 3

M oore JL, M o u q u e t N, L aw ton JH , L oreau M (2001) C oexis­ten ce , sa tu ra tio n a n d invasion re s is tan ce in sim u la ted p lan t a ssem blages. O ikos 94:303-314

M orrison LW (2000) M echan ism s of in terspecific com petition am ong invasive a n d tw o n a tiv e fire ants. O ikos 90:238-252

N aeem S, K nops JM H , T ilm an D, H ow e KM, K ennedy T, G ale S (2000) P lan t d iversity in c reases re s is tan ce to in vasion in th e ab sen ce of covary ing extrinsic factors. O ikos 91: 97-108

N o o n b u rg EG, S hu te r BJ, A bram s PA (2003) In d irec t effects of zeb ra m usse ls (D reissena po lym orpha) on th e p lan k to n ic food w eb. C an J F ish A q u a t Sei 60:1353-1368

R aw son PD, A g raw al V, H ilb ish T J (1999) H ybrid iza tion b e tw e e n th e b lu e m usse ls M ytilu s galloprovincialis an d Aí. trossulus a lo n g th e Pacific coast of N o rth A m erica: ev i­d en ce for lim ited in trog ression . M ar Biol 134:201-211

Rius M, M cQ uaid CD (2006) W ave action a n d com petitive in te rac tio n b e tw e e n th e invasive m usse l Aiy tilu s ga llo ­provincia lis a n d th e in d ig en o u s Perna p ern a in South Africa. M ar Biol 150 :69-78

S an ju an A, Z ap a ta C, A lvarez G (1997) G enetic d iffe ren tia ­tion in A iy tilu s galloprovincia lis Lmk. th ro u g h o u t th e w orld. O p h e lia 4 7 :13-31

Editorial responsib ility: R ichard Osm an,E dgew ater, A iaryland, USA

Sarver SK, Foltz DW (1993) G enetic p o p u la tio n s tru c tu re of a sp ec ie s ' com plex of b lu e m usse ls (R iytilus spp.). M ar Biol 117:105-112

► S h inen JL, M o rg an SG, C h an AL (2009) Invasion resis tan ce on ro ck y shores: d irec t a n d in d irec t effects of th re e n a tive p red a to rs on a n exotic a n d a n a tiv e p rey species. M ar Ecol P rog Ser 378:47-54

Stachow icz JJ , F ried H, O sm an RW, W hitlatch RB (2002) B iodiversity, invasion resis tan ce , a n d m arin e ecosystem function: reconciling p a tte rn a n d process. E cology 83: 2575-2590

Steffani CN, B ranch GM (2005) M echan ism s a n d co n se ­q u en ces of com petition b e tw e e n a n a lien m ussel, A iy tilu s galloprovincialis, a n d a n in d ig en o u s lim pet, Scutellastra argenvillei. J Exp M ar Biol Ecol 317:127-142

S to h lg ren TJ, B inkley D, C h o n g GW, K alk h an M A a n d o thers (1999) Exotic p lan t spec ies in v ad e ho t spots of n a tiv e p lan t diversity. Ecol M onogr 69 :25-46

S u ch an ek TH, G eller JB, K reiser BR, M itton JB (1997) Zoo- geo g rap h ic d istribu tions of th e sib ling spec ies R iy tilu s g a l­loprovincia lis a n d Aí. trossulus (Bivalvia: M ytilidae) an d th e ir hy b rid s in th e N o rth Pacific. Biol Bull 193:187-194

T h eb a u d C, F inzi A C, A ffre L, D ebussche M, E scarre J (1996) A ssessing w h y tw o in tro d u c ed C onyza differ in th e ir ab il­ity to in v ad e M e d ite rran e a n old fields. Ecology 77: 791-804

U nderw o o d A J (1986) T he analysis of com petition b y field experim ents. In; K ikkaw a J, A n d erso n D J (eds) C om m u­n ity ecology: p a tte rn a n d process. B lackw ell, M elbourne, p 240 -2 6 8

V an Erkom S churink C, Griffiths CL (1990) M arine m usse ls of so u th e rn Africa: th e ir d istribu tion p a tte rn s , s tan d in g stocks, exp lo itation a n d cu ltu re. J Shellfish Res 9 :75 -85

V elasco LA, N av arro JM (2005) F e ed in g physio logy of tw o b ivalves u n d e r lab o ra to ry a n d field conditions in re sp o n se to v a riab le food concen tra tions. M ar Ecol P rog Ser 291: 115-124

W ortham M J (2004) M ini-review : d istribu tion of th e M ed ite r­r a n e a n m usse l R iy tilu s galloprovincialis (Bivalvia: M ytili­dae) a n d hybrids in th e n o rth ea s t Pacific. J Shellfish Res 23:535-543

X avier BM, B ranch C M , W ieters E (2007) A b u n d an ce , g row th a n d re c ru itm en t of R iy tilu s galloprovincia lis on th e w est coast of Sou th A frica in re la tio n to u pw elling . M ar Ecol P rog Ser 346:189-201

Z ar JH (1974) B iostatistical analysis. P ren tice H aii, E n g le ­w ood Cliffs

Z ard i CI, N icastro KR, M cQ uaid CD, E rlandsson J (2008) San d a n d w av e in d u ce d m orta lity in invasive (R iy tilu s g a l­loprovincialis) an d in d ig en o u s (Perna perna) m ussels. M ar Biol 153:853-858

S u b m itted : J u n e 25, 2008: A ccep ted : F ebruary 17, 2009 Proofs re c e ive d fro m author(s): R ia y 8, 2009