MARINE ECOLOGY PROGRESS SERIES Vol. 167: 155-169, 1998 Published June 18 Mar Ecol Prog Ser
Spatial and temporal variability of patterns of colonization by mussels (Mytilus trossulus, M. edulis) on a wave-exposed rocky shore
Heather L. Hunt*, Robert E. Scheibling
Department of Biology. Dalhousie University, Halifax, Nova Scotia B3H 451. Canada
ABSTRACT: Colonization rates of mussels (Mytilus trossulus and M. edulis) were measured on natural substrata in tidepools and on emergent rock in recently ice-scoured and non-scoured regions of a rocky shore near Halifax, Nova Scotia, Canada. The relative importance of initial settlement/colonization, compared to subsequent dispersal and mortality, in determining the distribution and abundance of mussels was examined by comparing patterns and rates of mussel colonization at sampling intervals of days to months over a 17 mo period. Lrss than 4 ":, of mussels which colonized the quadrats sampled at short (2 to 7 d) intervals were settling larvae (
macrobenthic assemblaqe. the sllhc;tratl~m fnr co!oniza!ic~. 1:: addi:ion, thr- luny L~IIIIdbunaance oi colonists was linearly related to the cumulativct short term abundance during all but one of the inter- vals. Therefore, our results ~ndicatethat, over time scales up to 16 mo, patterns of initial colonization by settlers and larger post-larval mussels were inore important than post-colonizati.on mortality and dis- persal in determining pattt=rns of distribution and abundance of n~usselson this shore
KEY WORDS: Colonization Post-settlement dispersal . Mussels . lMytilus Ice scour . Intertidal zone Tidepools
INTRODUCTION (Woodin 1991). Post-settlement transport of juvenile molluscs can be actively initiated (Martel & Chia Studies of the role of recruitment variability in the 1991a), and is generally facilitated by the production of population and community dynamics of benthic long byssal or mucous threads which increase hydro- marine invertebrates have focused largely on pro- dynamic drag (Sigurdsson et al. 1976. De Blok & Tan- cesses influencing patterns of settlement by planktonic Maas 1977, Lane et al. 1985, Martel & Chia 1991a). larvae (reviewed by Butman 1987, Pawlik 2992, Young post-larval mussels can use threads to drift in Rodriguez et al. 1993).However, many benthic marine the water column at least until th~yreach a size of invertebrates, particularly molluscs, can disperse in -2 mm shell length (Sigurdsson et al. 1976, De Blok & the water column as juveniles (e.g. Sigurdsson et al. Tan-Maas 1977, Lane et al. 1985). Bayne (1964) 1976, Beukema & de Vlas 1989, Martel & Chia 1991b, demonstrated that Mytilus edulisentered a secondary Arrnonies 1992). When the rate of waterborne disper- pelagic phase at a size of 1 to 2 mm and moved from sal of juveniles is high, it may be an important deter- initial settlement sites on filamentous algae to a more minant of the distribution and abundance of adults permanent attachment on established beds of adult mussels. Although a temporary association between recently settled mussels and filamentous algae has 'Present address: MS 34, Woods Hole Oceanograph~clnstjtu- tion, Woods Hole, Massachusetts 02543, USA. been observed in several other studies (e.g. Seed 1969, E-mall: hlhunt@\uhoi.edu King et al. 1989), mussels also may settle directly onto
0 Inter-Research 1998 Resale of full article not perm~tted 156 Mar Ecol Prog Ser 167: 155-169, 1998
adult mussel beds (McGrath et al. 1988, Caceres- to compare the relative abundance of the 2 species Martinez et al. 1993, 1994). Patterns of post-settlement between habitats. In Ma.rch 1997, we collected 32 mus- dispersal may partially account for sporadic and un- sels in 3 size classes from each of 3 tidepools and 3 predictable pulses of recruitment which characterize adjacent areas of emergent rock. <5 mm (n = 8), many populations of Mytilus (Seed & Suchanek 1992). 5-9.9 mm (n = 8), and 10-24.9 mm shell length (SL) Although byssal drifting in the water column generally (n = 16). The mussels were identified to species using a involves small juveniles, larger mussels may be redis- polymerase chain reaction assay of a marker for inter- tributed by wave action or may move by crawling nal transcribed spacer regions between the 18s and (Paine 1974, Hunt 1997) Dispersal of large mussels 28s nuclear rDNA coding regions (Heath et al. 1995). may increase the rate of recovery of mussel cover after DNA was extracted from the whole animal for mussels disturbance and influence the dynamics of established 1964). We examined the effects of post-colonization We grouped the mussels collected at the shortest mortality and dispersal on patterns of initial coloniza- sampling intervals (2 to 7 d) into 4 size classes: RESULTS cantly between tidepools (76%) and emergent rock (66%) (G, = 1.20, p = 0.27). These results are consistent Species composition with a similar analysis of mussels collected the previ- ous year which indicated that -80 % of mussels in both We compared the frequency of Mytilus trossulus and habitats were M. trossulus (Hunt & Scheibling 1996). M. edulis (Table 1) among size classes (<5, 5-9.9, At another wave-exposed shore in Nova Scotia (ca 110 mm) and between tidepools and emergent rock 30 km away), Pedersen (1991) found that 77 to 91 D/oof using contin.gency tables. Hybrids were not included settlers ( Table 1 Frequency of Mytilus trossulus, M. edulis, and hybrids of the 2 species in 3 size classes (<5,5-9.9, and 10-24.9 mm) of musse1.s collected from tidepools and emergent rock Size class Habitat M, trossulus M. eduhs Hybrids Tidepool Emergent rock Tidepool Emergent rock Tidepool Emergent rock Total Tidepool Emergent rock Tldepools Non-scoured --- Scoured Emergent Rock -- Non-scoured - Scoured Fig. 1 Mytilus spp. Colonization rate of mussels of each of 4 size classes ( Table 2. Mytilus spp. Three-factor ANOVA of the cumulative density (no. 100cm-') of mussels of each of 4 size classes colonizillg quadrats sampled at short intervals from July to November 1993 and from May to November 1994. Factors are Habitat (H. tide- pools and emergent rock), Stratum (S, ice-scoured and non-scoured), and Plot (nested within H X S). Effects of Habitat. Stratum, and the interact~onbetween Habitat and Stratum were tested against the mean square error of Plot, and the effect of Plot was tested against the residual mean square error. 'p < 0.05; "p < 0.01; "'p < 0.001. The results of a posterior1 tests are presented in the 'Comparison' column. When there was a significant H X S interaction, means were compared between habitats within each stratum and between strata within each habitat uslng t-tests (TP = tidepool. ER = emergent rock. N-S = non-scoured, S = ice- scoured). = indicates a non-significant result, while < or > indicates a significant result (a= 0.05) Size class Source df MS P Comparison -- -- -P P 1993 0.5-1.9 mm Habitat 1 11.46 0.003 m TP: N-S=S Stratum 1 0 25 0 55 ER: N-S>S Habitat X Stratum 1 6.57 0.01 ' N-S: TP=ER Plot (H X S) 8 0 65 0.006 ' S: TP>ER Residual 24 0.18 2-4.9 mm Habitat 1 4.58 0.004 " TP: N-S=S Stratum 1 2 39 0.02' ER. N-S>S Habitat X Stratum 1 2.44 0.02' N-S. TP=ER Plot (H X S) 8 0.29 0.07 S: TP>ER Residual 24 0.14 >5 mm Habitat 1 2.45 <0.001"' TP: N-S=S Stratum 1 0.88 0.008" ER- N-S>S Habitat X Stratum 1 1.18 0.004. N-S. TP=EK Plot (H X S) 8 0.07 0.54 S: TP>ER Residual 24 0.08 1994 <0.5 mm Habitat 1 3 82 0.002" TP. N-S Stratum 1 2.06 0.009 m ER:W-S=S Habitat X Stratum 1 1.79 0.01 ' N-S: TP=ER Plot (H X S) 8 0.18 0.002 " S: TP>ER Residual 24 0.04 0.5-1 9 mm Habitat 1 6.26 0.004" TP: N-S=S Stratum 1 0.53 0.28 ER:N-S>S Habitat x Stratum 1 2.71 0.03' N-S: TP=ER Plot (H X S) 8 0.40 0.005" S: TP>ER Residual 24 0.10 2-4 9 mm Habitat 1 3.70 Stratum 1 0.06 Habitat X Stratum 1 1.18 Plot (Hx S] 8 0.24 Residual 24 0.10 >5 mm Habitat 1 4.60 <0.001 "' TP: N-S=S Stratum 1 0.12 0.40 ER N-S>S Habitat X Stratum 1 2.08 0.01 ' N-S: TP=ER Plot (H X S) 8 0.16 0.23 S: TP>ER Residual 24 0.110 were settling larvae (<0.5 mill SL). Settlement oc- August and October/November in 1994. Colonization curred from August to November, while larger mussels by larger size classes was at a peak when sampling colonized the quadrats over a more extended period began in July 1993 and peaked again in JuneIJuly from May to December. The daily colonization rate by 1994. Colonization by 0.5-1.9 mm mussels also peaked all size classes of mussels was negligible in winter/ in September/October in both years. early spring (January to April 1994). Settlement of ANOVA (Table 2) showed a significant interaction in ~0.5mm mussels peaked in September in 1993 and in the effects of Habitat and Stratum on the cumulative 160 Mar Ecol Prog Ser 167. 155-169, 1998 Scoured Non-scoured July-November 1993 0.5-1.9mm 2-4.9mm >5 mm Total 15 lo~c~m~ 202515 300400500 0 Fig. 2. ~Vytilus spp. 0 Mean (+ SE) cumula- 70 ::IA 10 200 tive number of mussels 5 50 0 5 100 (per 100 cm2) collected C from July to November c O 0 0 0 0 1993 and from May to .--U November 1994 from t Mav-November l994 the quadrats sampled >5 mm Total at short intervals in ice-scoured and non- scoured tidepools (TP) and emergent rock (ER). Data were aver- aged across 3 quadrats per plot and across 3 plots wi.thin each com- bination of habitat and stratum density of colonists for each size class of mussels in emergent rock (Fig. 2). We did not compare coloniza- both years, except for 2-4.9 mm mussels in 1994 when tion between years because sampling in 1993 began in the interaction was marginally non-significant and the July when colonization by large mussels was at a peak. cumulative density of colonists was significantly We examined the size distributions of mussel higher in tidepools than on emergent rock. Post hoc colonists by pooling individuals over the main colo- comparisons of means showed that settlers (<0.5 mm) nization periods in 1993 and 1994 (Fig. 3). Small mus- were 1 to 2 orders of magnitude more abundant in sels were proportionately more abundant in 1993, tidepools than on emergent rock in ice-scoured areas, when sampling began in July, than in 1994, when sam- but did not differ significantly between habitats in non- pling was carried out throughout the entire coloniza- scoured areas (Fig. 2, Table 2). Cumulative abundance tion period. In each year, 53% of colonists in any of settlers was significantly higher in ice-scoured than habitatktratum combination were recent settlers in non-scoured areas in tidepools, but did not differ (<0.5 mm). Mussels that were 0.5-1.9 mm represented significantly between strata on emergent rock. In con- the majority (50 to 81 %) of colonists in 1993, while trast, cumulative abundance of larger colonists was 2-4.9 mm mussels were the most abundant (53 to more similar among ice-scoured tidepools, and non- 71 %) colonists in 1994. Mussels that were >5 mm rep- scoured tidepools and emergent rock, but was consis- resented 3 to 8% of colonists in 1993 and 11 to 16% in tently lowest on ice-scoured emergent rock (Fig. 2). 1994. For each year, we compared the size distribu- Cumulative colonization by these larger mussels was tions of mussels between strata within each habitat significantly higher in tidepools than on emergent rock and between habitats within each strata using Kol- in ice-scoured areas, but did not differ significantly mogorov-Smirnov tests (Seigel & Castellan 1988) between habitats in non-scoured areas (Table 2). (Fig. 3). Mussels were pooled across quadrats and plots Moreover, colonization by larger mussels was signifi- within a habitat/stratum combination. In both years, cantly higher in non-scoured than in ice-scoured areas mussels difference was statistically significant in Tidepools Emergent Rock 1994, which reflected larger sample Non-scoured Scoured Non-scoured Scoured sizes in 1994 since the degree of differ- July-NOV93 ence remained small (D,,, = 0.07). Colonization over long sampling intervals The pattern of long term colonization was similar to the pattern of cumulative short term colonization (summed over all size classes of mussels): abundance was May-Nov 94 highest in ice-scoured tidepools and -0.15 lowest on ice-scoured emergent rock "'v (Fig. 4). For the 5, 11, and 16 mo inter- vals beginning in July 1993, ANOVA showed a significant interaction in the effects of Habitat and Stratum on the density of mussel colonists (Table 3). Post hoc comparison of means showed that the patterns of differences between habitats within strata, and between strata within habitats, were the same as Shell Length (mm) those of cumulative short term coloniza- tion of mussels >0.5 mm, i.e. coloniza- Fig. 3. Mytilus spp. Size frequency distributions of mussels collected from tion was greater in tidepools than on July to November 1993 and from May to November 1994 from the quadrats sampled at short intervals. Mussels were pooled across quadrats and plots emergent rock in ire-scoured areas, and ..-;.L;-v.,tL.A,L 2 ,,,,,,;,.a?:sr.rnm~,n of habit=.! cc! stratum. !n the first har. the hlack shad- greater on non-scoured than on ice- ing indicates settlers (~0.5mm SL), while the gray shad~ngindicates individ- scoured emergent rock. Colonization uals which were 0.5-0.9 mm. Brackets show results of Kolmogorov-Smirnov increased significantly over time, but not tests comparing the size distributions of mussels between strata within each habitat and between habitats within each stratum (D,,,, p; "'p > 0.5. 'p < 0.05, in proportion to the time elapsed: abun- "p < 0.01, "'p < 0.001) dance after 5 and 11 mo was 70 and I1 mo (Jul 93-May 94) .-V - -m Tidepool Emergent ~ock 6 mo 5 mo 500 (NO" 93-May 94) (May-Oct 94) Fig. 4. Mytilus spp. Mean (+ SE) den- Scoured 400 sity of mussel colonists in ice-scoured Non-scoured and non-scoured tidepools and emer- 300 gent rock after the 5, 11, and 16 mo intervals beginning in July 1993, and after the second and third 5 to 6 mo intervals. Data were averaged across 3 quadrats per plot and across 3 plots within each combination of habitat Tidepool Emergent Tidepool Emergent and stratum Rock Rock 162 Mar Ecol Prog Ser 167: 155-169, 1998 95%,, respectively, of the abundance after 16 mo (159 a significant interaction in the effects of habitat and mussels per 100 cm2, averaged across all habitathtra- stratum on the density of mussel colonists (Table 4). tum combinations). There were no significant interac- Differences between habitats within strata, and tions between time interval and either habitat or stra- between strata within habitats, were the same as for tum, nor was there a significant 3-way interaction. The the 5, 11, and 16 mo intervals beginning in 1993, density of colonists also varied significantly among except for the third 5 to 6 mo interval when the differ- plots within combinations of Habitat and Stratum. For ence between non-scoured and ice-scoured areas each of the three 5 to 6 mo intervals, ANOVA showed on emergent rock was marginally non-significant Table 3. ~Mytilusspp. ANOVA of the number of mussel colonists in 100 cm2 quadrats after the 5, 11, and 16 mo intervals begin- nlng in July 1993. Factors are Habitat (H, tidepools and emergent rock), Stratum (S, i.ce-scoured and non-scoured), Plot (nested w~thinH X S), and Time (5, 11, and 16 mo). Effects of Habitat, Stratum, and the interaction between Hab~tatand Stratum were tested agalnst the mean square error of Plot, and effects of Time and the interactions between Time, Habitat, and Stratum were tested against the mean square error of the interaction between Time and Plot. The effect of Plot and the interaction between Time and Plot were tested against the residual mean square error. 'p < 0.05; "p < 0.01; "'p c 0.001. The results of a posteriori tests are presented .in the 'Comparison' column. Since there was a significant H X S interaction, means were compared between habitats with~neach stratum and between strata withln each habitat using t-tests (TP = tidepool, ER = emergent rock, N-S =non- scoured, S = ~ce-scoured).Slnce there was a signiflcant effect of Time, means were compared among time intervals (5 = 5 mo, 11 = 11 mo. 16 = 16 mo) using SNK. = indicates a non-significant result, while < or > indicates a significant result (a= 0.05) Source P Comparison Habltat 0.001 " TP: N-S=S Stratum 0.22 ER: N-S>S Habitat X Stratum 0.001 N-S: TP=ER Plot (H X S) <0.001 "' S: TP>ER Tlme 0.004 5 < 11=16 Hab~tatX T~me 0.13 Stratum X Time 0.16 Habitat x Stratum X Time 0.57 Tlrne X Plot (H X S) 0.23 Residual Table 4. Myt~lusspp. ANOVA of the number of mussel colonists in 100 cm2 quadrats after th.e three 5 to 6 mo intervals. Factors are Habitat (H, ttdepools and emergent rock), Stratum (S, ~ce-scouredand non-scoured), and Plot (nested cv~thlnH X S). Effects of Habitat, Stratum, and the interaction between Habitat and Stratum were tested against the mean square error of Plot, and the effect of Plot was tested against the residual mean square error 'p < 0.05, "p < 0.01; "'p < 0.001 The results of aposterion tests are presented in the 'Comparison' column. When there was a significant H X S interaction, means wcrc? compared between habi- tats within each stratum and between strata within each habitat using t-tests (TP = tidepool, EK = emergent rock, N-S = non-scoured, S = ice-scoured). = ind~catesa non-slgnlficant result, while < or > indicates a significant result (a = 0.05) Interval Source Comparison .p 1st 5 to 6 mo Ha b~tat TP: N-S=S Stratum ER: N-S>S Habitat X Stratum N-S: TP=ER Plot (H x S) S: TP>ER Residual 2nd 5 to 6 mo Hab~tat TP: N-S=S Stratum ER: N-S>S Habltat X Stratum N-S: TP=ER Plot (H X S) S: TP>ER Residual 3rd 5 to 6 mo Habitat 1 TP: N-S=S Stratum 1 ER: N-S=S Habltat X Strdtum 1 N-S: TP=ER Plot (H X S) 8 S: TP>ER Hunt 8. Scheibling: Var~ab~l~tyof mussel colonization patlerns 163 Tidepools Emergent Rock After each of the longer term (5 to 16 mo) Non-scoured Scoured Non-scoured Scoured intervals, mussels 15 mm represented 72 to 5 mo (Jul-Nov 93) 100'2: of colonists and mussels > 10 mm repre- 10.13 n- I 0.18 ns~sented <10% (Fig. 5). We compared the size distributions of colonists between habitats within strata and between strata within habi- 40 tats for each interval. Differences in the size 20 distributions between combinations of habi- tat and stratum were not consistent over time and did not reflect the pattern observed in the 11 rno (Jul93-May 94) 019' I size distributions of cumulative short term 10.07' 80 10 15 "1 10.19 1 colonization (Figs. 3 & 5). However, after each of the 5 to 6 mo intervals, and the 11 and I 1 I I 16 mo intervals, mussels l mm SL were pro- portionately less abundant on ice-scoured emergent rock than in the other combinations of habitat and stratum. The absence of signif- 16 rno (Jul 93-0ct 94) 10.23 icant differences in size distributions in some 1 0 15 of the comparisons with ice-scoured emer- - 80 10.10"7 70.15 "1 "1 gent rock resulted from the smaller sample I l sizes from This habitat/stratum combination and not from smaller differences between cumulative size frequency distributions. We did not statistically compare the size distribu- 6 mo (Nov 93-May 94) tions among time intervals because the differ- -0.22 ns 0.28" I ences were clear and larger than the differ- 70.45"1 ences between habitats and strata within an 60,801 -0"01' ! i~terva!, and thc ar,a!yses ;vould have neces- - sitated a large number of comparisons. Mus- sels Comparison of sampling frequencies Shell Length (mm) To explore the importance of post-coloniza- tion mortality and dispersal in determining Fig. 5. Mytilus spp. Size frequency distributions of mussels after the 5, the abundance of colonists at the end of the 5 11, and 16 mo intervals beginning in July 1993, and the second and third to 16 mo intervals, we used regression analy- 5 to 6 mo ~ntervals.Mussels were pooled across quadrats and plots sis to examine the relationship between long within a combination of habitat and stratum. In the first bar, the black shading indicates settlers ( 5 rno 16 rno (Jul 93-Oct 94) 1 Fig. 6. Myfilus spp. Relationship between abundance of colonists (no 100cm-~)after a 5 to 16 mo interval and 5 mo the cumulative short (May-Oct 94) term colonization (no. Tidepools o Non-scoured 100cm-9 during that Scoured interval for the 5, 11, Emergent Rock and 16 mo intervals Non-scoured Scoured beginning in July 1993, I and the second and third 5 to 6 mo inter- vals. Data are averages for 3 quadrats per plot. The dashed line is the l-to-l line, and the solid line is the regres- Cumulative Short Term Colonization sion line term colonization also was significantly related to frequencies (D,,, = 0.25, p > 0.10). In each habitatktra- cumulative short term colonization, and the slope of tum combination during the third 5 to 6 mo interval, the relationship approached 1. However, during the mussels Scoured Tide~ools pools. Because many of the colonists were associated with macroalgae, our low esti- mates of mussel cover may partly reflect an artifact of our sampling method which recorded only the species/group in the top plane if there were several organisms at the same point on the substratum. How- ever, the relatively small size and low abundance (c3 cm-2) of mussel colonists at the end of each of the 5 to 6 mo inter- vals, the 11 mo interval, and the 16 mo Scoured Emergent Rock Non-scoured Emergent Rock interval indicates that the cover of mussels was probably not greatly underestimated. DISCUSSION Size of colonists and temporal pattern of colonization - Sampling at short intervals (2 to 7 d) Sept Apr Jul Oct Sept Apr Jul Oct indicated that more than 96% of mussels 93 94 94 94 93 94 94 94 colonizing the natural substratum were too large (20.5 mm) to be settling larvae. B Sheets Jointed calcareous Most of these mussels (67 to 82 % in 1994) t3 Filarnentous Crustose 6l Coarsely branched Barnacles were > 2 mm and some were as large as 20 0 Thick leathery Bare substratum to 25 mm. Because mussels of this size are prcbably too heavy to drift in the watei- Fig 7 Percentage cover of 6 functional form groups of macroalgae and of column using threads (Sigurdsson et al. barnacles and unoccupied substratum in ice-scoured and non-scoured tide- 1976, De Blok & Tan-Maas 1977, Lane et pools and emergent rock in September 1993 and April, August and October 1994. Data were averaged across short and long term colonization quadrats al. 1985), their redistribution occurs either within a plot (12 per plot in September 1993 and April 1994, 9 per plot in by dislodgment and deposition by waves, August and October 1994) and across 3 plots within each combination of or by crawling (Paine 1974, Hunt 1997). In habitat and stratum. Cover of Mytilus spp. in the long term colonization a tagging study, Hunt (1997) found that quadrats was not included and percentage cover of macroalgae and barna- mussels larger than 5 mm SL usually cles in these quadrats was adjusted to total 100% moved <5 cm over 4 wk. Because we cleared a 10 cm border around our sam- tose macroalgae than non-scoured emergent rock and pling quadrats, it is unlikely that postlarval colonists tidepools. In ice-scoured habitats, the macrobenthic entered the quadrats by crawling. The size range of assemblage differed between years: there wa.s a colonizing mussels encompassed most of the size greater cover of thick leathery algae and barnacles on range of the mussel population at Cranberry Cove: ice-scoured emergent rock, and of coarsely branched -90 % of mussels are smaller than 5 mm and <5% are algae in ice-scoured tidepools, and consequently less larger than 20 mm (Hunt & Scheibling 1995, in press). unoccupied space in October 1994 than in September Since growth rates are low (20.4 mm mo-' for mussels 1993. >5 mm; Hunt 1997), the large size of many of the For the quadrats sampled for long term colonization, colonists indicates that they were several years old. we also estimated the cover of mussels. Percentage The distance of wave dispersal of large mussels is cover of mussels was less than 1 % after each of the unknown. However, even relatively small-scale dis- three 5 to 6 mo intervals, with the exception of ice- persal should increase the rate of recovery of mussel scoured tidepools after the third 5 to 6 mo interval (F = cover from small-scale disturbances and affect the 6 %). On average, percentage cover was less than 3 % dynamics of established mussel aggregations. At the after 11 mo in each combination of habitat and stratum. same field site, Hunt (1997) found that changes in mus- After 16 mo, percentage cover of mussels was < 1 % on sel patch size were affected more by dislodgment and non-scoured and ice-scoured emergent rock, 8% in redistribution of mussels by waves than by mussel non-scoured tidepools, and 17 "/;, in ice-scoured tide- mortality. 166 Mar Ecol Prog Ser 167: 155-169, 1998 Sampling at short intervals indicated that the colo- Although spatial distribution is known to vary among nization rate of mussels >2 mm SL peaked in June to some of the closely related species in the M. edulis spe- July in both 1993 and 1994. In contrast, settlement cies complex (e.g. M. edulis and M. galloprovincialis; occurred in August to October, similar to the pattern Skibinski et al. 1983, Gosllng & McGrath 1990), varia- observed by Pedersen (1991) at another wave-exposed tion in the distribution of M. trossulus and M. edulis shore -30 km from our site. The peak of colonization of probably did not contribute to the pattern we ob- large mussels in late spring and summer may be asso- served, since genetic analysis indicated that the fre- ciated with the low wave action at this time of year. In quency of occurrence of the 2 species did not differ June and July of both years, able variability in the abundance of colonists among increase in the estimate of settlement in quadrats from quadrats within plots, most likely resulting from varla- which individuals have not been cleared is contrary to tion in the macrobenthic assemblage. the observations for barnacles (Connell 1985, Bertness The patterns of colonization on natural substrata cor- et al. 1992, Minchinton & Scheibling 1993a, b). How- respond to patterns of distribution and abundance of ever, barnacles require free space on the substratum mussels on ice-scoured but not on non-scoured regions for settlement, wheareas mussels generally settle on of the shore. Colonization rate of ice-scoured emergent filamentous or rough substrata, which can include the rock was low during all sampling intervals and mussels byssal threads of conspecifics. For example, Hunt are rare in these areas. In non-scoured areas, the spa- (1997) found a positive relationship between recruit- tial distribution and abundance of mussels differ be- ment rate of mussels to mussel patches and patch area. tween tidepools and emergent rock, but we did not The lack of a significant relationship between long detect a difference in colonization rate between the term and cumulative short term colonization for the 2 habitats. However, the small size and low percentage third 5 to 6 mo interval suggests that density-depen- cover of mussels after 16 mo (517%) suggests that dent post-colonization processes were more important differences in the mussel assemblage between the tide- during this interval, when cumulative short term colo- pools and emergent rock develop slowly. This is consis- nization was highest, than during the other 5 to 16 mo tent with studies of community recovery on wave- intervals. Post-colonization processes also influenced exposed shores on the Atlantic coast of Nova Scotia the size distributions of colonists. On ice-scoured after a large scale ice scour in 1987 (McCook & Chap- emergent rock (except in fall 1994), the size range of man 1997, Minchinton et al. 1997). On emergent rockin long term colonists was more limited than that of the mid intertidal zone, mussel cover was c20 % 14 mo colonists sampled at short intervals. This suggests that after the ice scour (McCook & Chapman 1997, Mlnchin- ice-scoured emergent rock is not a suitable substratum ton et al. 1997).The pattern of distribution and cover of for some sizes of mussels, particularly those Acknowledgements. We thank Norman Countway, Allan lapillus (L.)) predation on mussel (~Vytilus trossulus Hennigar, Anna Metaxas, Kristina Sander, and Alyson Shaw (Gould), M edulis (L.))assemblages in tidepools and on for assistance in the field. We are also grateful to Sharlenc emergent rock on a wave-exposed rocky shore in Nova Anthony for counting samples and to Dr S. Walde for Scotia, Canada. J Exp Mar Biol Ecol comments on an earlier draft of this manuscript. H.L.H. was K~ngPA, McGrath D, Britton W (1990) The use of artificial supported by a Natural Science and Engineer~ngResearch substrates in monitoring mussel (iLlytilus eduljs L.) settle- Council (NSERC) Postgraduate Award and an lzaak Walton ment on an exposed rocky shore in the west of Ireland. J Killam Memorial Scholarship from Dalhousie University, and Mar Biol Assoc LrK 70:371-380 the research was supported by an NSERC Research Grant to King PA, McGrath D, Gosling EM (1989) Reproduction and R.E.S. settlement of Mjftilus edulis on an exposed shore in Gal- way Bay, west coast of Ireland. J Mar Biol Assoc UK 69: 355-365 LITERATURE CITED Lane DJW, Beaumont AR. Hunter JR (1985) Byssus drifting and the drifting threads of the young post-larval mussel Armonies W (1992) M~gratory rhythms of drifting juvenile Mytilus edulis. Mar Biol 84:301-308 molluscs in tidal waters of the Wadden Sea. Mar Ecol Prog Littler MM (1980) Morphological form and photosynthetic Ser 83:197-206 performances of manne macroalgae: tests of a func- Bayne BL (1964) Primary and secondary settlement in Mytilus tional/form hypothesis. Bot Mar 22: 161-165 edulis L (Mollusca). J Anim Ecol33:513-523 Littler MM, Littler DS (1984) Relationships between macroal- Bayne BL (1965) Growth and the delay of metamorphosis of gal functional form groups and substrata stability in a sub- the larvae of Myt~lusedulis (L.).Ophelia 2.1-47 tropical rocky-intertidal system. J Exp Mar Biol Ecol 74: Bertness MD, Gaines SD, Stephens EG, Yund PO (1992) Com- 13-34 ponents of recruitment in populations of the acorn barna- Martel A. Chia FS (1991a) Foot-raising behaviour and active cle Semibalanus balanoides (Linnaeus). J Exp Mar Biol participation during the initial phase of post-metamorphic ECO~156 199-215 drifting in the gastropod Lacuna spp. Mar Ecol Prog Ser Beukema JJ, de Vlas J (1989) T~dal-current transport of 72.247-254 thread-drifting postlarval juveniles of the bivalve Macoma Martel A. Chia FS (1991b) Drifting and dispersal of small balthica from the Wadden Sea to the North Sea. Mar Ecol bivalves and gastropods with direct development. J Exp Prog Ser 52:193-200 Mar Biol .Ecol 150:131-147 Butman CA (1987) Larval settlement of soft sed~mentinverte- McCook LJ, Chapman ARO (1997) Patterns and variations in brates: the spat~alscales of pattern explained by actlve natura:l succession follocving massive ~ce-scourof a rocky habitat selection and the emerging role of hydrodynami- intertidal seashore. J Exp Mar Biol Ecol214:121-147 cal processes. Oceanogr Mar Biol Annu Rev 25:113-165 McGrath D, King PA, Gosling EM (1988) Evidence for the Caceres-Martinez J, Robledo JAF, Fi.gueras A (1993) Settle- direct settlement of Mytilus edulis larvae on adult mussel ment of mussels Mytilus gallopror~incialison an exposed beds. Mar Ecol Prog Ser 47:103-106 rocky shore in Ria de Vigo, NW Spain. marEcol Prog Ser McGuiness KA, Davis AR (1989) Analys~sand interpretation 93:195-198 of the recruit-settler relationship. J Exp Mar Biol Ecol 134: Caceres-Martinez J, Robledo JAF, Figueras A (1994) Settle- 197-202 ment and post-larvae behaviour of Mytilus galloprovin- Metaxas A, Hunt HL, Scheibling RE (1994) Spatial and tem- ciabs- field and laboratory experiments. Mar Ecol Prog Ser poral, vanability of mdcrobenthic cornmun.ities in tldepools 112:107-117 on a rocky shore in Nova Scotia, Canada. Mar Ecol Prog Connell JH (1985) The consequences of variation in initial set- Ser 105:89-103 tlement vs. post-settlement mortality in rocky intertidal Minchinton TE, Scheibling RE (1993a) Free space availability communities. J Exp Mar Biol Ecol 93:ll-45 and larval substratum selection as determinants of barna- De Blok JW, Geelen HJFM (1958) The substratum requlred rlr population structure in a developing rocky intertidal for the settling of mussels (Mytilus edulis L.). Arch Neerl community. Mar Ecol Prog Ser 95:233-244 2001 13:446-460 Minchinton TE, Scheibling RE (1993b) Variations in sampling De Rlok JW, Tan-Maas M (1977) Function of byssus threads In procedure and frequency affect estimates of recruitment young postlarval Mytilus. Nature 267:558 of barnacles. Mar Ecol Prog Ser 99:83-88 Gosling EM, McGrath D (1990) Genetic variability in Minchinton TE, Scheibling RE, Hunt HL (1997) Recovery of exposed-shore mussels, Mytilus spp., along an environ- an intertidal assemblage following massive ice scour in mental gradient. Mar Biol 104:413-418 Nova Scotia. Canada. Bot Mar 40:139-148 Heath DD, Racvson PD, Hilbish T (1995) PCR-based nuclear Palne RT (1974) Intertidal community structure: experimental markers identify alien blue mussel (Myt~lusspp.) geno- studies on the relationship between a dominant competl- types on the west coast of Canada. Can J Fish Aquat Sci tor and its principal predator. Oecologia 15:93-120 52:2621-2627 Pawlik JR (1992) Chemical ecology of the settlement of ben- Hunt HL (1997) Structure and dynamics of intertidal mussel thic marine invertebrates. Oceanogr Mar Biol Annu Rev (Mytllus trossulus, M. edulls) assemblages. PhD thes~s, 30:273-335 Dalhousie University, Halifax Pedersen Eh1 (1991) Population genetics of Mytllus from a Hunt HL. Scheibling RE (1995) Structure and dynamics of developing intertidal community in Nova Scotia, Canada. mussel patches in tidepools on a rocky shore in Nova Sco- MSc thesis, Llalhousie University, Halifax tia, Canada. Mar Ecol Prog Ser 124:105-115 Price HA (1980) Seasonal variation in the strength of byssal Hunt HL, Scheibling RE (1996) Physical and biological factors attachment of the common mussel Mytilus edulls L. J Mar influencing mussel (Myfilus trossulus, M. edulis) settle- Biol Assoc UK 60: 1035-1037 ment on a wave-exposed rocky shore. Mar Ecol Prog Ser Price HA (1982) An analysis of factors determining seasonal 142:135-l45 variation in the byssal attachment strength of Myt~lus Hunt HL, Scheibling RE (in press) Effects of whelk (Nucella eduUs. J Mar Biol Assoc UK 62:147- 155 Hunt & Scheibling- Variability of mussel colonizat~onpatterns l69 Rodriguez SR, Ojeda FP, Inestrosa NC (1993) Settlement of the behavloral sciences. hlcGraw-Hill, New York benthic marine invertebrates. Mar Ecol Prog Ser 97 S~gurdssonJB, Titman C\Y, Dav~esPA (1976)The dispersal of 193-207 young post-larval bivalve molluscs by byssus threads. Seed R (1969) The ecology of Mytilus edulis L. (Lamelli- Nature 262:386-387 branchiata) on exposed rocky shores I. Breeding and set- Sklbinskl DOF, Beardmore JA, C~ossTF (1983)Aspects of the tlement. Oecologia 3:2?7-316 population genetics of ,Vvt~lus(hlytllidae; Mollusca) in the Seed R,Suchanek TH (1992)Population and community ecol- Brltlsh Isles Blol J Linn Soc 19.137-183 ogy of Mytilus. In: Gosling .E (ed) The musscl Mytilus- Witman JD, Suchanek TH (1984) Musscls in flow: drag and ecology, physiology, genetics, and culture. Elsevier, New d~slodgementby eplzoans. M~I-Ecol Prog Ser 16:259-268 York, p 87-169 Llloodin SA (1991) Recruitment of Infauna: positive or nega- Selgel S, Cdstelldn NJ Jr (1988) Nonparametrlc statist~csfor tive cues? Am Zoo1 31 797-807 Editorial responsrbility: Roger Hughes (Contributing Editor), Submitted: December 2, 1997; Accepted: March 18, 1998 Gwynedd, UK Proofs received from author[s): May 19, 1998