Distribution of Mytilus edulis and M. trossulus on the Gaspe Coast in relation to spatial scale.

ABSTRACT Mytilus edulis and M. trossulus are the two molluscan species co-occurring along the Gaspe coast, eastern Canada. Few studies examine the effects of environmental factors (temperature, salinity, tidal height and wave action) on the local or regional distribution of these two species in areas where they co-occur. This study is designed precisely to examine the niche of each of the two species, particularly in relation to environmental factors in the northeastern part of the Gulf of St. Lawrence, Canada. Mussels were sampled on rocky shores according to a factorial design including six locations, 3 degrees of wave exposure (exposed, semi exposed and sheltered) and 2 intertidal levels (mid and low). Species were distinguished with polymerase chain reaction amplification using a diagnostic DNA marker (Glu-5). Relative frequencies of each species showed no clear patterns of distribution with wave exposure or tidal height. A pattern of distribution at the regional scale does occur, but this pattern could not be related to salinity or temperature gradients observed.

KEY WORDS: Mytilus edulis, Mytilus trossulus, distribution, wave action, tidal height

INTRODUCTION

Mussels in the genus Mytilus have been the subject of some controversy since the discovery of M. trossulus using genetic markers in the 1970s (Gosling 1992), in particular in relation to their taxonomic status at the specie level. Initially, classification of individuals within the Mytilus complex was based on morphologic characters. Phenotypic plasticity leads to the classification of several species and subspecies within the genus (Gosling 1992). Recently, the study of polymorphic loci by protein electrophoresis revealed the existence of three genetically distinct groups (Koehn et al. 1984, McDonald & Koehn 1988, McDonald et al. 1991). It was proposed that M. edulis, M. galloprovincialis and M. trossulus should be considered as 3 distinct species (Koehn et al. 1984, McDonald & Koehn 1988, McDonald et al. 1991, Varvio et al. 1988). Thus, leading to the recognition of 4 species in the genus Mytilus complex: M. edulis, M. trossulus, M. galloprovincialis and M. californianus (Gosling 1992).

M. edulis and M. trossulus are the 2 species occurring in the northwest Atlantic Ocean (Gosling 1992). In Newfoundland, species identification by allozyme electrophoresis showed mixed populations at nearly all sites with an overall prevalence of M. edulis (Penney & Hart 1999). High frequencies of M. edulis were also found in the southern part of Nova Scotia and in the Bay of Fundy (Mallet & Carver 1999). The coasts of Prince Edward Island, Northumberland Strait and Magdalen Islands were characterized by pure or almost pure M. edulis populations (Varvio et al. 1988). In contrast, pure populations of M. trossulus were observed in Bras d'Or Lakes (Cape Breton) (Mallet & Carver 1995, Mallet & Carver 1999) and on the Lower North Shore (Quebec) of the Gulf of St. Lawrence (Thomas et al. 2002).

Previous sampling campaigns along the shores of the Gaspe Peninsula showed higher relative frequencies of M. trossulus in Miguasha, Gaspe Bay and Riviere-au-Renard (see Fig. 1). Populations dominated by M. edulis were located in the New Richmond area (see Fig. 1; Thomas et al. 2002). This pattern intuitively suggests that the distribution of the two species of mussels is driven by environmental spatial variability within the Baie des Chaleurs. This work was carded out to determine the factors responsible for the local or regional distribution of the two species coexisting in the Baie des Chaleurs and the Gasp6 Peninsula. We hypothesize that the distribution of M. edulis and M. trossulus is influenced by temperature, salinity and wave action gradients observed at different spatial scales where higher frequencies of M. trossulus are expected in cold and low salinity waters.

[FIGURE 1 OMITTED]

METHODS

Study Zone

Baie des Chaleurs is a partly mixed estuarine system opening into the Gulf of St. Lawrence. The Restigouche River is the major tributary located at the head of the bay. Other less important rivers also flow in the Bay. The Gaspe current is also a strong hydrographic feature influencing the study area. This current develops near Rimouski and flows seaward along the north shore of the Gaspe Peninsula until it reaches Cape Gaspe where it spreads southeasterly offshore and becomes weaker (El-Sabh 1976). Thus, current direction inside the Baie des Chaleurs is oriented to the west on the north shore and oriented to the east on the south shore, creating a cyclonic circulation inside the bay (Legendre & Watt 1970, Tamigneaux 1996). Surface water temperature forms a positive gradient from north to south and from east to west ends of the bay (Habbane et al. 1997). At the head of the bay, surface water temperature can reach 18[degrees]C during summer. Salinity can fluctuate roughly from 27.5 [per thousand] to 30 [per thousand] (Habbane et al. 1997).

Genetic Analysis

Species identification was done by polymerase chain reaction (PCR) using the diagnostic loci marker glu-5. This method allows distinguishing between the two species and hybrid individuals. It is a more accurate method than allozyme analysis and requires little biologic material. Thus, small juvenile individuals can be easily analyzed (Rawson et al. 1996). For mussel DNA isolation, we used QIAGEN DNeasy Tissue Kit and 50 ng of DNA served as PCR template.

Sampling

In October 1999, a preliminary sampling program was carried out. Six stations (Riviere-au-Renard, Gaspe, Grande-Riviere, Port-Daniel, New Richmond, Escuminac) (Fig. 1) located along the Gaspe peninsula were chosen according to the following criteria: presence of an embayment, natural rocky shores and wave action gradient (exposed, semi exposed and sheltered sites) associated with the embayment.

Experiments followed a factorial design where mussel samples were collected at 2 tidal levels (lower and mid intertidal levels) of 3 levels of wave exposure (exposed, semi exposed and sheltered sites) at each station mentioned earlier. A total of 36 samples each containing 30 juvenile mussels were collected between October 19 to 25, 1999 and their species identified. The mussels measured between 2.10-18.63 mm in shell length (mean shell length = 7.65 [+ or -] 3.26 mm), as determined using a caliper ([+ or -] 0.01 mm). Mussels were assumed to be from the same class age, but probably from different spawning events. Adjacent sites were separated by at least 200 m and stations by ~80 km. Mussels were haphazardly sampled within the mussel beds. A few samples (sheltered sites of New Richmond, Escuminac and Gaspe) had to be collected on soft substrata. The same sampling design was repeated from September 25 to 30, 2000, except for samples from soft bottom. In 2000, 50 juveniles per sample (0.80-14.48 mm; mean shell length = 3.37 [+ or -] 1.89 mm), were collected.

Environmental Measurements

In 2000 only, water temperature, salinity and wave force were measured once a week, at each site, during low tide periods from June 28 to August 24. Each station was visited on the same day of the week. Water samples were collected at the same tidal phase (low tide) and brought to the laboratory for salinity measurements. Two dynamometers identical to those of Bell and Denny (1994) were attached to boulders near lower water level at each site. The dynamometers precalibrated allowed to estimate the strongest wave force, each week.

Statistical analyses were performed with SAS v 8.0 (SAS Institute) using Genmod. Preliminary analyses on hybrid individuals were performed, but the number of hybrid individuals was too low to calculate any model. Thus, data analysis was performed on frequencies of M. edulis using a binomial distribution. The Genmod model analyzed M. edulis frequencies using all possible interactions with spatial factors (station, site and tidal level) included in the factorial design. Environmental data (temperature, salinity and wave action) were used to establish a relation with the observed patterns and were used for the multiple regression.

RESULTS

Species Distribution in 1999

In 1999, sample size varied from 24 to 41 mussels (n total = 1067). All samples from station of Escuminac were dominated by M. trossulus (Fig. 2) with relative frequencies of M. edulis varying between 35% and 42%. The population at station of New Richmond showed the highest relative frequencies of M. edulis (frequencies between 51% to 87%). Proportions of M. edulis were variable among sites and tidal level at stations of Port-Daniel and Grande-Riviere, where relative frequencies of M. edulis varied respectively from 25% to 70% and 16% to 60%. Finally, M. trossulus was the predominant species at stations Gasp6 and Riviere-au-Renard in 1999. Populations of M. edulis varied from 14% to 46% at station Gaspe and from 7% to 38% at station Riviere-au-Renard.

[FIGURE 2 OMITTED]

Statistical analysis on the frequencies of hybrids observed in 1999 failed to calculate estimates for the model due to low frequencies and high number of samples containing no hybrid individuals. Thus, statistical analyses were performed only on M. edulis and the patterns observed were considered complementary to those of M. trossulus. Genmod model on M. edulis frequencies showed a significant triple interaction (P = 0.0432; Table 1). Multiple comparisons using sequential Bonferroni correction (Rice 1989; Table 2) with gradual decreasing relative frequencies of M. edulis were carried out. No significant interaction between the low and the mid intertidal levels were observed except for a few samples. No significant differences in the relative frequencies of M. edulis were observed among sites at the two intertidal levels at station Escuminac. Similarly, no distribution trend was observed in the frequencies of M. edulis at the different sites of the two intertidal levels or from one station to another at the same intertidal level, even if some significant differences (4 significant differences/18 total multiple comparisons) were observed.

Relation to Environmental Factors

During summer 2000, environmental data were collected. Surface water temperature formed a west-east gradient from Escuminac where mean water temperature was 19.2[degrees]C ([+ or -] 2.4) to Riviere-au-Renard where mean water surface temperature was 13.5[degrees]C ([+ or -] 2.9). Differences in mean water temperature between adjacent stations varied about 1[degrees]C to 2[degrees]C. During the observation period, the coldest water temperatures recorded at each site generally occurred on first day of measurement (end of June). In contrast, the warmest water temperatures were measured at the end of July and the beginning of August (17[degrees]C at Riviere-au-Renard; 25[degrees]C at Escuminac and New Richmond). Salinity varied from 20 to 25 ppt and showed no specific west-east gradients (Fig. 3A). The semi exposed site at Escuminac, the sheltered sites at Grande-Riviere and Riviere-au-Renard were located near a fresh water source and, accordingly, relatively low salinities were observed at these stations (Fig. 3A). Wave action (Fig. 3B) was higher at exposed than at sheltered sites at stations New Richmond where values of mean wave force in the sheltered site was 6.6 N ([+ or -] 5.4 N) whereas values of the mean wave force at the semi exposed and the exposed sites were somewhat similar 11.8 ([+ or -] 4.1 N) and 11.5 N ([+ or -] 4.6 N). The same pattern is repeated at stations Grande-Riviere and Gaspe where values of mean wave force of the semi exposed sites were similar to the values obtained at the sheltered sites with values around 9.5 N ([+ or -] 4.5 and 2.6 N) at station Grande-Riviere and around 6.5 N ([+ or -] 1.4 and 1.7 N) at station Gasp6. At these stations, values of the mean maximum wave force at the exposed sites were respectively of 15.9 N ([+ or -] 7.5 N) at station Grande-Riviere and 10.2 N ([+ or -] 2.6 N) at station Gaspe. Finally, the stations Escuminac, Port-Daniel and Riviere-au-Renard show no particular maximum wave exposure gradient between the three different wave exposure degrees. Mean values of maximum wave force at station Escuminac varies between 5.0 N ([+ or -] 3.6 N) in the exposed site and 6.4 N ([+ or -] 5.7 N) at the semi-exposed site. Mean values of maximum wave force at station Port-Daniel varied from 5.5 N ([+ or -] 1.0 N) at the semi-exposed site and 8.5 N ([+ or -] 2.4 N) at the exposed site. Values of mean maximum wave force at station Riviere-au-Renard varied from 4.2 N ([+ or -] 0.8 N) at the semi-exposed site and 7.2 N ([+ or -] 2.9 N) at the exposed site.

[FIGURE 3 OMITTED]

Number of mussels per sample for the year 2000 varied between 30 and 50 (n total = 1504). Again, relative frequencies of M. edulis at station Escuminac varied between 31% and 43% except for the semi-exposed site where it reached 66% (Fig. 4). New Richmond was the station that showed the highest frequencies of M. edulis (56% to 86%). Proportions of the two species at Port-Daniel were variable, but M. edulis generally dominated the population with frequencies ranging from 48% to 64%. In contrast to 1999, M. trossulus were more frequent at Grande-Riviere (frequencies of M. edulis from 28% to 48%) in 2000. Gaspe showed higher frequencies of M. edulis in 2000 than in 1999. The sheltered and semi exposed sites were characterized by proportions of M. edulis over 50% but the proportions at the exposed site were lower than 30%. The population of Riviere-au-Renard was characterized by high frequencies of M. trossulus, from 57% to 71%, except at the semi-exposed site where the proportion was 35%.

[FIGURE 4 OMITTED]

Statistical analysis on these frequencies using Genmod (SAS Institute, Table 3) detected 2 significant interactions: Station X site (P = 0.0025) and Site X level (P = 0.0001). Multiple comparisons examining the effects of station and site (Table 4) showed no significant difference in frequencies of M. edulis due to wave exposure at all station, except at Gasp6 where the frequency of M. edulis at the exposed site was significantly lower than that observed at the sheltered or the semi-exposed site. When the stations were compared at the different wave exposures, station New Richmond was generally grouped with station Port-Daniel except at the exposed site. These stations showed the highest frequencies of M. edulis. Stations Escuminac, Riviere-au-Renard and Grande-Riviere were grouped together at the three wave exposure levels except for the sheltered site of Riviere-au-Renard, where the sample was missing. These samples were characterized by high frequencies of M. trossulus. In the interaction Site X tidal level there was a significant difference between frequencies at different tidal levels at the semi-exposed sites, but this difference was not observed at the exposed sites. Multiple comparisons were not carried out between low and mid intertidal levels at the sheltered sites because of the missing sample at station Riviere-au-Renard. Also, there were no significant differences at the three different sites at the low intertidal level. There was a significant difference between the frequencies at semi-exposed and exposed sites in the mid-intertidal level where higher frequencies of M. edulis were observed at the semi-exposed site. The sheltered site is not shown because of missing data.

Given that there was discrepancy between the anticipated maximum wave force and the estimated wave force, we performed a multiple regression on frequencies of M. edulis with environmental data (mean water surface temperature, mean salinity and mean wave force). Two variables were entered in the model: mean water surface temperature and salinity (P > 0.0218; model R-square: 0.2126; M. edulis frequencies = -24.56 + 2.90 mean water surface temperature + 1.17 salinity). It is interesting that maximum wave force was not retained by the model, but this result was expected, given that there were no coherent distribution patterns with respect to this factor (Table 2, Fig. 4B).

DISCUSSION

Our results show no differences in frequencies between low and mid intertidal levels at most of the compared sites. In 1999, 4 sites out of 18 shared a significant difference between the two tidal levels; 3 of the 4 sites had higher frequencies of M. edulis in the low intertidal level (Table 2) than in the mid intertidal level. In 2000, there was a significant difference between the frequencies of M. edulis, but only at the semi-exposed sites where high frequencies of M. edulis were observed in the mid intertidal level (Table 4B). Thus, no general pattern of distribution in relation with the tidal height can be drawn for the two species found in the Baie des Chaleurs and the Gaspe Peninsula.

Hybridizing between M. edulis and M. trossulus on North American Atlantic coast is less extensive than between M. edulis and M. galloprovincialis in Europe (Innes & Bates 1999). Frequencies of hybrid mussels in the Canadian maritime provinces were either very low in Nova Scotia (Mallet & Carver 1995, Mallet & Carver 1999) and in Magdalen Islands (Tremblay et al. 1998) or null in Newfoundland (Bates & Innes 1995) and Prince Edward Island (Mallet & Carver 1999). However, mussels identified by PCR with diagnostic loci markers (glu-5 or Its) showed 23% hybrid mussel frequencies in Nova Scotia and 26% in Newfoundland (Comesana et al. 1999). Thus, the variations observed in the level of hybridation may be related to the technique used: protein electrophoresis versus PCR. The combination of markers will increase the level of hybrid observed, but evidence for limited hybridization persists. Toro et al. (2004) have examined, with two PCR markers, the genetic composition of early life-history stages of mussels in a site of Newfoundland where M. edulis and M. trossulus co-occur and observed no evidence for differential morality. They concluded that the low frequency of hybrids observed is probably related to a combination of pre and post zygotic reproductive isolating mechanisms operating early in the mussel's life history. Laboratory experiments have demonstrated incompatibility between M. edulis and M. trossulus at the gamete stage (Rawson et al. 2003).

Within European hybrid populations of M. edulis and M. galloprovincialis, study of allozymes (Octopine dehydrogenase, Odh; Esterase, Est-D) on juveniles showed that allele frequencies of M. galloprovincialis were higher than in M. edulis from the intertidal zone (Gosling & McGrath 1990). It was proposed that differences in distribution in the tidal zone were due to differences in selection intensity (Gosling & McGrath 1990). In contrast, Gilg and Hilbish (2000) did not find differential primary or secondary settlement of M. edulis and M. galloprovincialis in relation to tidal level. They suggested that selection favoring M. gaUoprovincialis at high tidal level occurred after settlement (Gilg & Hilbish 2000). M. edulis distribution occurred higher on the shore than that of M. californianus (Suchanek 1978) but, no significant effects of shore level was observed for M. trossulus and M. galloprovincialis (Sarver & Foltz 1993) even if Hofman and Somero (1995) suggested that M. trossulus were more thermosensitive than M. galloprovincialis (Hofmann & Somero 1995). Clearly more work is needed on this topic.

Mussel survival is partly dependent on attachment to the substratum. The different Mytilus species exhibit different attachment strength: M. galloprovincialis > M. trossulus (Bell & Denny 1994) and M. californianus > M. edulis (Witman & Suchanek 1984). M. galloprovincialis > M. edulis at exposed compared the sheltered sites at one location (Gardner 1994), Sarver and coworker (1993) found no relation between the distribution of two Mytilus species (M. trossulus and M. galloprovincialis) and the wave action. In Newfoundland (Canada), one study reported higher frequencies of M. trossulus at the exposed sites at two locations, but the authors noted that further investigations were required on this topic (Bates & Innes 1995).

At the larger scale (>100 km), the relative frequencies of both species observed in 1999 and 2000 are in agreement with the results of Thomas et al. (2002), indicating that pure or almost pure populations of M. trossulus were observed in Riviere-au-Renard and Miguasha, and pure populations of M. edulis were located in New Richmond. In the 1999 sampling no pure populations were observed, but those of Riviere-an-Renard and Escuminac were dominated by M. trossulus; populations in New Richmond were dominated by M. edulis. Other stations showed mixed populations of the two species in variable frequencies. Temperatures recorded in 2000 apparently never reached lethal levels: larvae survive in water from 5[degrees]C to 25[degrees]C and optimal growth occurs at 20[degrees]C and 20-35 ppt (HRS-Brenko & Calabrese 1969). Intertidal organisms may experience thermal stress during emersion periods and metabolic costs combined with thermal stress could influence species distribution pattern (Hofmann & Somero 1995). Differential thermal sensitivity measured between M. trossulus and M. galloprovincialis by the synthesis of stress protein hsp-70 suggests that M. trossulus is the most sensitive to high temperatures (Hofmann & Somero 1996). Because we observed no particular patterns of distribution for each species in relation to shore level, we suppose that temperatures related to shore level were too small to induce differential survival.

M. trossulus populations generally dominate where salinity is reduced. Pure M. trossulus populations were found in the Baltic sea (4-5 ppt; Gosling 1992) or in Bras d'Or Lakes (<20 ppt; Mallet & Carver 1999). Qiu and coworkers (2001) tested the tolerance of the 2 species (M. edulis and M. trossulus) during larval development. At 15 ppt, no M. edulis larvae and few M. trossulus survived. At 20 ppt, survivorship of larvae was lower for M. edulis than for M. trossulus suggesting that M. edulis is less tolerant to reduced salinity during larval development (Qiu et al. 2001). Salinity recorded during year 2000 (Fig. 4B), reflect the presence of near freshwater sources in Escuminac (semi exposed), New Richmond (sheltered), Grande-Riviere (sheltered) and Riviere-au-Renard (sheltered). These sites are located near a river or a brook mouth that flows directly into bays. Salinity data were collected during low tide periods once a week but almost all other sites had salinity values over 20 ppt. As salinities over 20 ppt were observed at all sites and stations, selective pressures on larval development would not likely favor one species over the other. This might explain why M. edulis was observed in location know to harbor "pure populations" of M. trossulus, for instance at Riviere-au-Renard, Gaspe and Escuminac. We hypothesize that if selection occurs at low salinity, then it will occur when larvae are in their early developmental stages. Finally, temperature and salinity explained 21% of the variability of M. edulis frequencies with the multiple regressions performed on the 2000 data. Environmental factors such as salinity and, to a lesser degree, temperature have been shown to have a significant influence on the distribution of 2 species (M. trossulus and M. galloprovincialis) all along the coast of California (Sarver & Foltz 1993).

[FIGURE 4 OMITTED]

Our study cannot pinpoint the factors responsible for the local distribution of M. edulis and M. trossulus in the Baie des Chaleurs and the Gaspe Peninsula. Tidal amplitude, wave exposure and temperature and salinity gradients in the bay may not vary sufficiently to explain differential distribution between the two species.

[FIGURE 3 OMITTED]

ACKNOWLEDGMENTS

The authors thank E. Parent for his technical support and B. Tremblay, A. Moreau and R. Kervan for collecting data on the field. This research was supported by the Technopole Maritime (grant and scholarship) and GIROQ awarding to V. M. within the GIROQ research program.

LITERATURE CITED

Bates, J. A. & D. J. Innes. 1995. Genetic variation among populations of Mytilus spp. in eastern Newfoundland. Mar. Biol. 124:417-424.

Bell, E. C. & M. W. Denny. 1994. Quantifying "wave exposure": a simple device for recording maximum velocity and results of its use at several field sites. J. Exp. Mar. Biol. Ecol. 181:9-29.

Comesana, A. S., J. E. Toro, D. J. Innes & R. J. Thompson. 1999. A molecular approach to the ecology of a mussel (Mytilus edulis-M, trossulus) hybrid zone on the east coast of Newfoundland, Canada. Mar. Biol. 133:213-221.

El-Sabh, M. 1976. Surface circulation pattern in the Gulf of St. Lawrence. J. Fish. Res. Board. Can. 33:124-138.

Gardner, J. P. A. 1994. The structure and dynamics of naturally occurring hybrid Mytilus edulis Linnaeus, 1758 and Mytilus galloprovincialis Lamarck, 1819 (Bivalvia: Mollusca) populations: Review and interpretation. Arch. Hydrobiol. Supp. 99:37-71.

Gilg, M. R. & T. J. Hilbish. 2000. The relationship between allele frequency and tidal height in a mussel hybrid zone: a test of the differential settlement hypothesis. Mar. Biol. 137:371-378.

Gosling, E. M. 1992. Systematic and geographic distribution of Mytilus. In: E.M. Gosling, editor. The mussel Mytilus: ecology, physiology, genetics and culture, New York: Elsevier. pp. 1-17.

Gosling, E. M. & D. McGrath. 1990. Genetic variability in exposed-shore mussels, Mytilus spp., along an environmental gradient. Mar. Biol. 104:413-418.

Habbane, M., M. I. El-Sabh & J.-M. Dubois. 1997. Potentiel aquacole par teledetection passive et SIG: application aux eaux cotieres de la baie des Chaleurs (est du Canada). Int. J. Remote Sens. 18:3439-3457.

Hofmann, G. E. & G. N. Somero. 1995. Evidence for protein damage at environmental temperatures: Seasonal changes in levels of ubiquitin conjugates and hsp70 in the intertidal mussel Mytilus trossulus. J. Exp. Biol. 198:1509-1518.

Hofmann, G. E. & G. N. Somero. 1996. Interspecific variation in thermal denaturation of proteins in the congeneric mussels Mytilus trossulus and M. galloprovincialis: Evidence from the heat-shock response and protein ubiquitination. Mar. Biol. 126:65-75.

HRS-Brenko, M. H. & A. Calabrese. 1969. The combined effects of salinity and temperature on larvae of the mussel Mytilus edulis. Mar. Biol. 4:224-226.

Innes, D. J. & J. A. Bates. 1999. Morphological variation of Mytilus edulis and Mytilus trossulus in eastern Newfoundland. Mar. Biol. 133:691-699.

Koehn, R. K., J. G. Hall, D. J. Innes & A. J. Zera. 1984. Genetic differentiation of Mytilus in eastern North America. Mar. Biol. 79:117-126.

Legendre, L. & W. D. Watt. 1970. The distribution of primary production relative to a cyclonic gyre in Bale des Chaleurs. Mar. Biol. 7:167-170.

Mallet, A. L. & C. E. A. Carver. 1995, Comparative growth and survival patterns of Mytilus trossulus and Mytilus edulis in Atlantic Canada. Can. J. Fish. Aquat. Sci. 52:1873-1880.

Mallet, A. L. & C. E. A. Carver. 1999. Maritime distribution and commercial production performance of Mytilus edulis and Mytilus trossulus. 13. Aquac. Assoc. Can. 99:7-13.

McDonald, J. H. & R. K. Koehn. 1988. The mussels Mytilus galloprovincialis and Mytilus trossulus on the Pacific coast of North America. Mar. Biol. 99:111-118.

McDonald, J, H., R. Seed & R. K. Koehn. 1991. Allozymes and morphometric characters of three species of Mytilus in the northern and southern hemispheres. Mar. Biol. 111:323-333.

Penney, R. W. & M. J. Hart. 1999. Distribution, genetic structure, and morphometry of Mytilus edulis and M. trossulus within a mixed species zone. J. Shellfish Res. 18:367-374.

Qiu, J. W., R. Tremblay & E. Bourget. 2001. Ontogenic changes in hyposaline tolerance in the mussels Mytilus edulis and M. trossulus: implications for distribution. Mar. Ecol. Prog. Ser. 228:143-152.

Rawson, P. D., K. L. Joyner, K. Meetze & T. J. Hilbish. 1996. Evidence of intragenic recombination within a novel genetic marker that distinguishes mussels in the Mytilus edulis species complex. Heredity 77: 599-607.

Rawson, P. D., C. Slaughter & P. O. Yund. 2003. Patterns of gamete incompatibility between the blue mussels Mytilus edulis and M. trossulus. Mar. Biol. 143:317-325.

Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225.

Sarver, S. K. & D. W. Foltz. 1993. Genetic population structure of a species' complex of blue mussels (Mytilus spp.).Mar. Biol. 117:105-112.

Suchanek, T. H. 1978. The ecology of Mytilus edulis L. in exposed rocky intertidal communities. J. Exp. Mar. Biol. Ecol. 31:103-120.

Tamigneaux, E. 1996. Changements dans la structure de taille des assemblages de phytoplancton et consequences pour les reseaux trophiques pelagiques dans des eaux cotieres temperees. Ph.D. thesis, Universite Laval, Quebec, Canada. 158 pp.

Thomas, B., R. Tremblay & M.-P. Turcotte. 2002. Distribution de Mytilus edulis et de M. trossulus dans les regions maritimes de l'Est du Quebec et des Maritimes entre 1996-1998. Cahier d'information no. 138, Ministere de l'Agriculture, des Pecheries et de l'Alimentation. Direction de l'innovation et des technologies. 70 pp.

Toro, J., D. J. Innes & J. Thompson. 2004. Genetic variation among life-history stages of mussels in a Mytilus edulis-M, trossulus hybrid zone. Mar. Biol. 145:713-725.

Tremblay, R., B. Myrand & J.-M. Sevigny. 1998. Genetic characterization of wild and suspension-cultured blue mussels (Mytilus edulis Linneaus, 1758) in the Magdalen Islands (Southern Gulf of St. Lawrence, Canada). J. Shellfish Res. 17:1191-1202.

Varvio, S.-L., R. K. Koehn & R. Vainola. 1988. Evolutionary genetics of the Mytilus edulis complex in the North Atlantic region. Mar. Biol. 98:51-60.

Witman, J. D. & T. H. Suchanek. 1984. Mussels in flow: drag and dislodgment by epizonans. Mar. Ecol. Prog. Ser. 16:259-268.

VALERIE MOREAU, (1) REJEAN TREMBLAY (2) * AND EDWIN BOURGET (3)

(1) GIROQ, Universite Laval, Quebec, Canada, G1K 7P4; (2) Centre Aquicole Marin-UQAR, MAPAQ, 6 rue du Parc, C.P. 340, Grande-Riviere, Quebec, Canada, G0C 1VO and (3) Vice-rectorat a la recherche, Universite de Sherbrooke, Sherbrooke, Quebec, Canada, J1K 2R1

* Corresponding author. Present address: Dr Rejean Tremblay Institut des sciences de lamer, Universite du Quebec a Rimouski 310 allee des Ursulines Rimouski, Qc, Canada G5L 3A1; Tel.: [phone], p1705; Fax: [phone]; E-mail: [e-mail]

TABLE 1.
Results of Genmod analysis examining the effects of Station
(Escuminac, New Richmond, Port-Daniel, Grande-Riviere, Gaspe,
Riviere-au-Renard), Site (sheltered, semi-exposed, exposed) and
Level (low intertidal level, mid intertidal level) on the frequencies
of M. edulis from samples collected in 1999.

Source of Variation      df   [chi square]   Pr > [chi sqaure]

Station                   5      89.90           <0.0001 *
Site                      2       5.49            0.0642
Station X site           10      37.85           <0.0001 *
Level                     1       6.31            0.0120 *
Station X level           5       6.07            0.2990
Site X level              2       8.66            0.0132 *
Station X site X level    9      17.37            0.0432 *

Significant differences: * P [less than or equal to] 0.05.

TABLE 2

Results of multiple comparisons tests examining the effects of the
interaction between the factors Station, Site and Tidal level.

                        Decreasing Niveau
                           Within Site

Escuminac           sheltered      low   mid
                    semi-exposed   mid   low
                    exposed        mid   low
New Richmond        sheltered      low   mid
                    semi-exposed   mid   low
Port-Daniel         exposed        low   mid
                    sheltered      low   mid
                    semi-exposed   low   mid
                    exposed        mid   low
Grande-Riviere      sheltered      low   mid
                    semi-exposed   low   mid
                    exposed        mid   low
Gaspe               sheltered      missing
                    semi-exposed   low   mid
                    exposed        mid   low
Riviere-au-Renard   sheltered      low   mid
                    semi-exposed   low   mid
                    exposed        low   mid

                          Decreasing Site Within Niveau

Escuminac           low intertidal level   shel   semi   exp
                    mid intertidal level   exp    semi   shel
New Richmond        low intertidal level   shel   exp    semi
                    mid intertidal level   semi   shel   exp
Port-Daniel         low intertidal level   semi   exp    shel
                    mid intertidal level   exp    semi   shel
Grande-Riviere      low intertidal level   semi   shel   exp
                    mid intertidal level   semi   exp    shel
Gaspe               low intertidal level   exp    semi
                    mid intertidal level   shel   exp    semi
Riviere-au-Renard   low intertidal level   exp    shel   semi
                    mid intertidal level   semi   exp    steel

Effects connected by a line do not differ significantly from one
another with the sequential Bonferonni correction P = 0.05.

TABLE 3.
Results of the Genmod analysis examining the effects of Stations
(Escuminac, New Richmond, Port-Daniel, Grande-Riviere, Gaspe,
Riviere-au-Renard). Sites (sheltered, semi-exposed, exposed) and
Levels (low intertidal level, mid intertidal level) on the frequencies
of M. edulis from samples collected in 2000.

                               [chi     Pr > [chi
Source of Variation      Df   square]    square]

Station                   5    68.40    <0.0001 *
Site                      2     6.90     0.0317 *
Station X site           10    27.09     0.0025 *
Level                     1     2.70     0.1001
Station X level           5     6.56     0.2552
Site X level              2    18.19     0.0001 *
Station X site X level    9    12.61     0.1810

Significant differences: * P [less than or equal to] 0.05.

TABLE 4.
Multiple comparisons examining the effects of Stations and Sites B,
Multiple comparisons examining the effects of Site and Level.

(A)

Decreasing Site Within Station

Escuminac           semi   shel   exp
New Richmond        semi   exp    shel
Port-Daniel         semi   exp    shel
Grande-Riviere      exp    semi   shel
Gaspe               semi   shel   exp
Riviere-au-Renard   semi   exp

Decreasing Station Within Site

sheltered      NR   Ga   PD   E    GR
semi-exposed   NR   PD   Ga   E    RR   GR
exposed        NR   PD   GR   E    RR   Ga

(B)

Decreasing Level With Site

sheltered      missing
semi-exposed   mid       low
exposed        low       mid

Decreasing Site Within Level

low intertidal level   shel   exp   semi
mid intertidal level   semi   exp

Effects connected by a line do not differ significantly from one
another with the sequential Bonferonni correction at P = 0.05.