Temporal variation in the settlement of blue mussels (Mytilus edulis C. Linnaeus, 1758 and M. trossulus Gould 1850) in Eastern Maine.

ABSTRACT Blue mussels in the genus Mytilus are cultured extensively in the Northeastern United States and throughout much of Atlantic Canada. Recent reports suggest that of the two species that co-occur in this region, the economic value of Mytilus edulis is higher than that of its congener M. trossulus. Because mussel farmers typically rely on wild caught spat to scud grow-out facilities, the productivity of mussel farms may benefit from a method of spat collection that limits the frequency of M. trossulus spat. To explore the potential for developing such a method in eastern Maine, where M. trossulus is highly abundant, we investigated the spatial and temporal variation in the proportion of M. edulis and M. trossulus spat settling at three sites in Cobscook Bay, Maine, USA in 2000. We round little evidence of location- or depth-dependent variation in the frequency of M. edulis spat. There was a significantly greater proportion of M. edulis spat on the collectors retrieved in October and December relative to those retrieved in August and September. However. the low abundance of settlers in October and December, coupled with the likelihood that spat settling late in the year will experience low growth rates and high mortality as a function of rapidly declining water temperature and food supplies, suggests that there are significant impediments to the preferential collection of M. edulis seed within Cobscook Bay.

KEY WORDS: settlement, spat collection, mussel culture, spatial variation, temporal variation, Mytilus

INTRODUCTION

The blue mussel, Mytilus edulis (C. Linneaus 1758), is a common, rocky intertidal and subtidal bivalve mollusk, which is harvested extensively in the Northeastern United States and Atlantic provinces of Canada. Despite a growing market demand for blue mussels in the U.S., the domestic production of mussels has steadily declined from a peak of 4800 metric tons in 1988 to 1400 metric tons in 2001 (NMFS 2003). Among the states contributing to the U.S. production, the vast majority (~88%) are harvested in Maine. Historically mussels have been harvested from natural beds and bottom culture leases seeded with juvenile mussels. Recently, however, mussel culturists in Maine have adopted suspended raft techniques commonly used in parts of Europe and Atlantic Canada. Because off-bottom techniques use a much greater proportion of the water column they are much more efficient and produce a product considered to have superior meat yield and quality relative to that obtained from bottom culture (Lutz 1980, Hickman 1992, Roberts et al. 1999). Although the majority of suspended raft and other off-bottom culture operations are presently located on the central coast of Maine, potential suspended raft culture sites can be found throughout much of the state.

The presence of a second species of blue mussel, M. trossulus (Gould 1850), however, may present an impediment to the expansion of raft culture to sites in eastern Maine. Genetic surveys using allozyme electrophoresis and PCR-based molecular markers have shown that M. trossulus is widely distributed throughout the Canadian Maritime Provinces, often in mixed populations with M. edulis (e.g., Koehn et al. 1976, 1984, Bates & Innes 1995, Comesana et al. 1999, Innes & Bates 1999, Mallet & Carver 1999, Penney & Hart 1999, Penney et al. 2002). Recently, Rawson et al. (2001) found that blue mussel populations throughout much of eastern Maine contain relatively high frequencies of M. trossulus, as well. Freeman et al. (1994) have strewn that M. trossulus mussels obtain less meat than M. edulis mussels and are more prone to shell fracture during processing. In addition, given species-specific growth rates, Mallet and Carver (1995) have estimated that the economic value of M. edulis is 1.7 times that of M. trossulus under off-bottom culture conditions. Because raft culture facilities are typically seeded with wild caught mussel spat, the proportion of M. edulis and M. trossulus under cultivation is dependent on the relative frequency with which these species settle on spat collectors. Thus, in locations where M. edulis and M. trossulus are sympatric, a method of spat collection that favors M. edulis seed and limits the frequency of M. trossulus seed would be highly beneficial to the productivity of suspended culture facilities (Freeman 1996).

The development of such methods, in turn, is dependent upon there being sufficient temporal or spatial variation in species-specific settlement. Freeman et al. (2002) and Kenchington et al. (2002) explored the potential for the differential settlement of M. edulis and M. trossulus larvae at depth in mesocosm and field studies, respectively. The results of their mesocosm study suggest that species-specific larval behaviors result in the greater settlement of M. edulis at depth when there is no thermocline present. In their field studies, they examined the patterns of mussel settlement at two sites in Nova Scotia, Canada, throughout the spawning season in 3 successive years. Although the timing of species-specific settlement varied among years, they consistently observed a significantly higher percentage of M. edulis spat on collectors deployed 4-5 m below the surface. Based on their results, they recommended that mussel growers consider deploying subsurface spat collectors to increase the proportion of M. edulis postlarvae.

In this study, we investigated the spatial and temporal variation in the frequency of M. edulis and M. trossulus settlement at three sites in Cobscook Bay, Maine, USA to determine whether such variation will facilitate the development of spat collection methods favoring M. edulis and the expansion of raft culture operations to eastern Maine. Although our sampling strategy differs somewhat from that of Kenchington et al. (2002), we were particularly interested in whether the increased settlement of M. edulis with depth observed by Kenchington et al. (2002) is found elsewhere in the zone of sympatry between M. edulis and M. trossulus.

MATERIALS AND METHODS

Spat Collection

The temporal and spatial patterns of mussel settlement were investigated by placing spat collectors at three locations in Cobs cook Bay (Fig. 1). Cobscook Bay is an extremely dynamic embayment where the morphology of the bay in combination with a large tidal range (>8 m) creates strong directional flows within much of the bay (Brooks et al. 1999). We placed spat collectors at three sites that lie within the major flow patterns within the bay. All three sites were located 30-50 m from shore in approximately 35-40 m of water (at low tide). Collectors were suspended from the peripheral anchor lines and buoys of salmon net pens at Shackford Head (SH), from an equipment supply raft on a salmon aqua culture site in South Bay (SB), and from the anchor lines of an experimental scallop farm in East Bay (EB). Spat collectors consisted of 6 m lengths of aged and frayed 13-mm thick polypropylene rope, also known as "pot warp". Each collector was suspended from the surface and weighted at the bottom using cinder blocks (12.3 kg) to keep the collector vertical in the water column. Three replicate collectors were initially deployed at each site on June 27, 2000 and recovered on August 8 at which time a second set of collectors was deployed. Collectors were recovered and redeployed on September 28 and October 28, and the experiment was terminated when the last set was retrieved on December 28. Upon recovery, the spat collectors were cut into three 1.5-m sections (top, middle, and bottom) starting from the surface. Each section was placed in an individual Ziploc bag containing 95% ethanol, and transported on ice to the University of Maine where they were stored at 4[degrees]C. Subsequently, each rope section was removed from the 95% ethanol, the rope strands were unwound, and all spat gently removed. These spat were combined with any that had fallen off and settled to the bottom of the Ziploc bag, counted by hand, placed in 50-ml centrifuge tubes containing fresh 95% ethanol, and stored at 4[degrees]C.

[FIGURE 1 OMITTED]

Molecular Analysis

Previous studies investigating genetic variation in northwest Atlantic blue mussel populations have used various combinations of allozyme and PCR-based markers to identify M. edulis and M. trossulus individuals and their hybrids (e.g., Koehn et al. 1976, 1984, McDonald et al. 1991, Bates & Innes 1995, Comesana et al. 1999). Because PCR-based methods require very little tissue they are ideally suited to studies involving larvae or recently settled postlarvae. Thus, we estimated the proportion of M. edulis and M. trossulus among the spat on each section of rope using two PCR-based genetic assays. Template DNA was prepared from approximately 20 individual spat from each rope section by placing whole spat in a 0.5-mL microcentrifuge tube containing 20 ng of proteinase K in 10 [micro]L of a buffer containing 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100 and incubating at 55[degrees]C for 3 h on a rotating platform. The samples were heated to 95[degrees]C for 15 min before 10 [micro]L of GeneReleaser (BioVentures Corp.) was added. The resulting mixtures were incubated at 65[degrees]C for 30 s, 8[degrees]C for 30 s, 65[degrees]C for 90 s, 97[degrees]C for 180 s, 8[degrees]C for 60 s, 65[degrees]C for 180 s, 97[degrees]C for 60 s, and finally 65[degrees]C for 60 s, after which the samples were centrifuged at full speed (13,200 rpm) for 2 min. The supernatant containing DNA was transferred to a new 0.5-mL microcentrifuge tube and was stored at 4[degrees]C. The isolated DNA was used in the PCR-based Glu-5' and ITS genetic assays that are considered diagnostic for M. edulis and M. trossulus. The application of these markers followed the protocols described in Rawson et al. (1996) and Heath et al. (1995), respectively.

Statistical Analysis

Variation in the combined settlement of M. edulis and M. trossulus with respect to site, date of collection, and depth was analyzed by a 3-way full factorial fixed effects (model I) analysis of variance (ANOVA). Of the 36 ropes deployed in this study, four were lost prior to recovery. Although no more than one rope was lost at any single location during any sampling period, this creates imbalance in our dataset so that the appropriate F statistics for the main effects and all interaction terms were estimated using Type III sums of squares. To eliminate heteroscedasticity among the residuals, total counts of spat abundance were log transformed prior to analysis. The validity of the model and data transformation was further determined by testing whether the residuals were normally distributed using a Levene's test.

Variation in the frequency of M. edulis Glu5' and ITS alleles among the spat recovered from individual rope segments was also analyzed using ANOVA. For many of the ropes retrieved in August, October, and December, we were unable to find and genetically characterize more than a couple (<5) spat per rope segment (Table 1). This created a severe imbalance in our dataset and precluded the use of a 3-way ANOVA to simultaneously analyze site- and depth-specific variation in the frequency of M. edulis alleles across all four sampling periods. Thus, to analyze temporal variation in the abundance of M. edulis spat, we first calculated the M. edulis Glu5' allele frequency among the spat collected and genotyped at each site within each month, without regard for individual ropes or depths. The site-specific allele frequencies were then analyzed in a 1-way ANOVA with sampling date as the independent variable. One of our main goals in this study is to investigate whether species-specific settlement varies as a function of depth. In this regard, we obtained sufficient allele frequency data to analyze the depth- and site-specific variation in the frequency of the M.edulis Glu5' allele for the ropes collected in September using a 2-way fixed effects ANOVA. In both analyses, the frequency of the M. edulis Glu-5' allele was normalized using an arcsin-transformation. The appropriate F-statistics for the main effects and any interaction terms were constructed using Type III sums of squares and the validity of the model and data transformation determined by graphical analysis of the homogeneity of the residuals and testing whether the residuals were normally distributed using a Levene's test. An identical approach was also used to examine the degree to which variation in the frequency of the M. edulis ITS allele was related to location, depth, or sampling date.

RESULTS

We observed substantial spatial and temporal variation in the settlement of blue mussels, M. edulis and M. trossulus, in Cobscook Bay. The total abundance of mussel spat on rope segments varied from 0 on collectors retrieved from East Bay in December to nearly 1,500 per 1.5 m rope segment at Shackford Head in September (Fig. 2). The 3-way ANOVA model explained over 90% of the variance in settlement in our data set (adjusted [R.sup.2] = 0.90; F = 15.06; P < 0.001) and identified a significant interaction involving all three of the main effects in the model (Table 2). This interaction is primarily due to a shift from a positive association between the abundance of settlers and depth in August and December to a negative association in September and October at Shackford Head, and only a weak dependence of spat abundance on depth at the other two sites (see Fig. 2). Two of the main effects in this model, month and site, were also significant. In general, blue mussel settlement was much higher at Shackford Head compared with both the East Bay and South Bay sites, up to 5-fold higher in the September and October collections. Peak settlement occurred in September at all three locations, whereas much lower settlement was observed on the collectors retrieved in October and December and virtually no spat were found on the collectors recovered in August.

[FIGURE 2 OMITTED]

We isolated DNA from a total of 1316 spat sampled from 48 rope segments and successfully genotyped 1249 (95%) of these individuals at the Glu5' marker. Our success rate for the ITS marker was lower and only 1022 spat were successfully genotyped at this marker. Of the 1020 individuals that successfully genotyped both markers, the allele frequency variation at the Glu5' marker was highly correlated with the variation at the ITS marker across sampling date, site, and depth (r = 0.95; P < 0.001). Because the statistical analyses of the variation at the Glu5' and ITS markers are quantitatively and qualitatively similar and the higher success rate for the Glu5' marker, we have chosen to present only the ANOVA results pertaining to the spatial and temporal variation in the frequency of the M. edulis Glu5' allele.

The frequency of the M. edulis Glu5' allele was fairly consistent within collection dates and across depths and the majority of the variation in M. edulis Glu5' allele frequency occurred between sampling dates (Fig. 3). Unfortunately, the lack of settlers on many of our collectors precluded a formal statistical analysis of allele frequency variation with respect to depth at all three sites across all tour collection dates. However, we pooled the frequency data across sites and depths and used a 1-way ANOVA to investigate whether there was significant temporal variation in the frequency of the M. edulis Glu5' allele. This model was highly significant (adjusted [R.sup.2] = 0.95; F = 41.00; P < 0.001; Table 3) and post hoc tests. (Tukey's) of means indicated that significant differences in the average allele frequency of the M. edulis Glu5'allele occurred between sampling dates. This significant difference is due to an increase in the frequency of the M. edulis Glu5' allele from approximately 40% on ropes retrieved in August and September to 80% on the ropes retrieved in October and December (Fig. 3 & 4). In addition, we obtained genotypic data from enough of our collectors in September to allow us to use a 2-way ANOVA to investigate the degree to which the frequency of the M. edulis Glu5' allele varies as a function of sampling location and depth. This ANOVA model explained only 43% of the variation in the arcsin-transformed allele frequency for the collectors retrieved from all three sites in September (adjusted [R.sup.2] = 0.43; F - 1.21; P = 0.37; Table 4). The low [R.sup.2] value indicates that most of the variation in M. edulis Glu5' allele frequency occurred among individual rope segments. Not surprisingly, neither the 2-way interaction nor the main effects were significant in this analysis, although it should be noted that this ANOVA had relatively little power to detect such effects (see Table 3).

[FIGURES 3-4 OMITTED]

We used the information from both of our genetic markers to estimate the frequency of individuals with M. edulis, M. trossulus, and mixed di-locus genotypes. Boecklen and Howard (1997) have shown that at least 4-5 well-differentiated genetic markers are necessary to reasonably estimate the frequency of backcross hybrid genotypes. We successfully genotyped 1200 mussel spat at two diagnostic markers. Though this is too few markers for estimating introgressed genotypes, we can use our genotypic data to roughly classify individuals into M. edulis, M. trossulus and mixed or hybrid genotypes. The frequency of M. trossulus and hybrid genotypes both peaked in September and then declined to around 12% and 4%, respectively, in our October and December samplings (Fig. 5). As expected from the shift in the frequency of the M. edulis Glu5' allele frequency, M. edulis genotypes were predominant on the collectors retrieved in October and December. Finally, we used the genotype frequency data to estimate the abundance of M. edulis spat at the Shackford Head site as a function of sampling date and depth. The highest abundance of settlers was observed at Shackford Head; however, given the temporal shift in genotype frequencies, the abundance of M. edulis settlers remains relatively constant through time (Table 5)

[FIGURE 5 OMITTED]

DISCUSSION

The mussel culture industry in Maine currently uses wild caught spat to seed both bottom and off-bottom culture sites. In central Maine, the majority of mussel seed for off-bottom culture originate from the Blue Hill Salt Pond (see Fig. 1). The expense and logistics of seed transfer between central and eastern Maine, as well as concern regarding the potential for disease transfer associated the movement of shellfish between rivers and bays in Maine make it highly desirable to identify a source of mussel seed in eastern Maine. The studies of Freeman (1994) and Mallet and Carver (1995, 1999), however, indicate that the productivity of off-bottom culture operations can be substantially decreased when there is an increased frequency of M. trossulus under culture. Given the prevalence of M. trossulus adults in Cobscook Bay and well into the Gulf of Maine (Rawson et al. 2001), capitalizing on local sources of seed requires the development of spat collection methods that limit the frequency of M. trossulus. For this reason. we investigated species-specific patterns of settlement among the mussel spat settling on our collectors in Cobscook Bay.

Larval settlement in marine benthic invertebrates, particularly those with an extended period of planktonic development, is the result of an interplay between various factors including local and regional hydrodynamics (Pineda 2000, Sponaugle et al. 2002). The morphology of Cobscook Bay together with the large tidal range in the region (>8 m) creates strong directional flows in the main channel of the bay (Brooks et al. 1999). As these tidal flows enter the center of the bay, they in turn create an eddy dipole consisting of one clockwise gyre in the southern portion of East Bay and a counter-clockwise gyre in the northern portion of South Bay (Brooks et al. 1999).

Our collectors were deployed near these hydrodynamic features and we anticipated large numbers of mussel postlarvae at all three of our sites. However, we observed substantial variation in mussel settlement among sites. The relatively low settlement on the collectors at our South Bay site was surprising, given that these collectors were placed at a salmon net pen facility within the main current associated with the South Bay gyre. Heavy fouling of the net pens by blue mussels is a continuing problem at this facility (C. Bartlett, personal observation). We can only surmise that because our collectors were downstream of the net pens they were essentially in a settlement shadow. Likewise, the low settlement at our East Bay site suggests a diminished supply of postlarvae. In this case, we feel that our collectors were located further north of the East Bay gyre then we had initially intended.

In contrast, we observed much higher settlement at our Shackford Head site. The value of sites like Shackford Head as a local source of mussel spat within Cobscook Bay, however, is dependent upon the degree to which settlement for M. edulis and M. trossulus is temporally or spatially segregated. We found no evidence of location-specific variation in the relative abundance of M. edulis and M. trossulus spat on our collectors. This observation is consistent with our previous observation that M. trossulus adults are highly abundant throughout Cobscook Bay (Rawson el al. 2001). Blue mussels have a planktonic development period that typically lasts 30 days or more and likely results in an open population structure providing little opportunity for spatial variation in the abundance of M. trossulus or M. edulis larvae to develop.

We found little evidence of depth specific variation in the abundance of M. edulis spat, as well. Low settlement on many of our collectors prevented us from analyzing the effect of depth on the frequency of M. edulis markers at all sampling dates. However, depth had no effect on either the frequency of the Glu5' nor the ITS M. edulis allele in September when spat abundance on our collectors peaked. The maximum difference between the average depth-specific frequency of the M. edulis Glu5' allele was 0.06%. Although the ANOVA testing for depth-specific variation had low statistical power, post hoc analysis suggests that even if we had deployed five times as many collectors as we did in this study, the power to detect such small differences would still have been <0.3.

The lack of depth-specific variation in the abundance of M. edulis spat was unanticipated, given the recent report by Kenchington et al. (2002). They examined the species-specific patterns of blue mussel settlement at two sites in Nova Scotia and observed consistently higher settlement of M. edulis on collectors 4-5 m below the surface. The discrepancy between our study and that of Kenchington et al. (2002) may be due to the different sampling strategies used in each. Kenchington et al. (2002) genotyped mussel spat from collector segments that were 0-1m, 2-3 m, and 4-5 m below the surface. Thus, our sampling strategy did not extend as deep (only 4.5 m) and our sampling intervals were not as discrete as those of Kenchington et al. (2002). It is interesting to note, however, that Kenchington et al. (2002) observed an increase in the frequency of M. edulis spat even at 2-3 m below the surface at their Parrang Cove site (see Fig. 3 in Kenchington et al. 2002). A similar increase in the frequency of M. edulis postlarvae below 3 m in Cobscook Bay should have been detected by our methods.

A significant shift in the frequency of M. edulis Glu5' and ITS alleles was apparent across sampling dates in our study. The average frequency of these alleles increased from 40% to 45% in August and September to 75% to 80% in October and December. This shift was unexpected, given the overlapping reproductive cycles for M. edulis and M. trossulus in Cobscook Bay. Maloy et al. (2003) examined the seasonal changes in gamete volume fraction and oocyte diameter for an intertidal population of blue mussels located near our East Bay collectors. They found that maximal gonadal development for both species occurred in mid July followed by a spawning event in late July and early August of 2000. Assuming a 30-d period of larval development, the timing of the spawning event observed by Maloy et al. (2003) correlates well with the peak settlement we observed at our three sites in September. Similarly, Toro et al. (2002) observed overlapping reproductive cycles for M. edulis and M. trossulus in a sympatric population in Newfoundland. They also noted a protracted period of spawning activity for M. trossulus and hybrid mussels. The increase in the frequency of M. edulis spat in October and December in our study, however, would suggest a more protracted spawning season for M. edulis adults in Cobscook Bay. Interestingly, Maloy et al. (2003) noted that there was a larger, though not statistically significant, decline in gamete volume fraction for M. edulis relative to M. trossulus in early September, consistent with a second spawning event for the former species. In addition, histologic analysis on the same specimens indicated that the highest proportion of individuals with undifferentiated gonads occurred from September to December for M. trossulus whereas few undifferentiated M. edulis were observed until January (Maloy personal communication). These two observations suggest that M. edulis is more gametogenically active during the period between September and December in Cobscook Bay. The variation in the frequency of hybrid mussels we observed is also consistent with species-specific temporal variation in spawning activity. The sharp decline in the frequency of M. trossulus and hybrid genotypes in our October and December samplings suggests that M. edulis continues to spawn after most M. trossulus individuals have ceased spawning.

Alternatively, differential mortality in the plankton or during metamorphosis may result in the shift in genotype frequencies we observed. A third potential hypothesis is that the increase in the frequency of M. edulis spat on our collectors in October and December is the result of secondary settlement. The former hypothesis requires that both M. trossulus and hybrid larvae experience much higher levels of mortality compared with M. edulis during October and November whereas the latter hypothesis requires that there be substantial genotype-specific behavioral differences among recently settled mussels. Presently, there is little comparative data on the larval physiology or behavior of M. edulis, M. trossulus, and their hybrids, so at present we favor the hypothesis that species-specific variation in spawning activity produces the temporal shift to higher frequencies of M. edulis spat.

Finally, our data have important implications for the potential establishment of off-bottom culture of blue mussels in Cobscook Bay. The results we have presented suggest that the deployment of spat collectors from October through December could result in >80% M. edulis spat. Three important caveats, however, accompany this observation. First, our sampling was confined to a single year and without additional sampling it is impossible to tell whether the increase in M. edulis spat late in the year will hold in other years. Second, the overall abundance of spat decreased dramatically on the collectors we retrieved in October and December so that even Shackford Head, the best of our three sites, may not receive enough spat to seed commercially viable culture facilities. Third, spat setting later in the year are likely to realize low growth rates and potentially high mortality due to declining water temperatures. Given these caveats, reliance on spat collected late in the year would appear to be a risky strategy. The consistently high frequency of M. trossulus spat at all three sites and depths in September, when peak settlement occurs, would seem to preclude the development of site- or depth-specific methods for the collection of M. edulis spat as well as the deployment of M. edulis seed from external sources up to October. Thus, our results indicate that there are likely to be significant obstacles limiting the preferential collection and cultivation of M. edulis mussels in Cobscook Bay.

TABLE 1.
Number of rope segments (spat collectors) for which [greater
than or equal to] 10 spat were retrieved and genetically
identified during the 4 sampling periods in this study.
Three depths were sampled with T, M, and B corresponding
to the top, middle, and bottom 1.5 m rope segments,
respectively.

                   August    September    October    December

                 T   M   B   T   M   B   T   M   B   T   M   B

Shackford Head   0   2   0   3   3   3   2   2   2   1   2   2
South Bay        0   0   0   3   2   3   1   1   1   0   2   1
East Bay         1   1   2   2   2   1   1   1   1   0   0   0

TABLE 2.
Three-way ANOVA (full model) of log-transformed settlement
frequency for blue mussel spat at 3 sites in Cobscook Bay
during 2000.

     Source            df     SS     [F.sub.ratio]     P

Month                   3   105.82       85.44       0.000
Site                    2    42.35       51.29       0.000
Depth                   2     2.75        3.31       0.053
Month * Site            6    35.61       14.38       0.000
Month * Depth           6    14.02        5.66       0.000
Site * Depth            4     2.19        1.32       0.272
Month * Site * Depth   12    10.66        2.15       0.027
Residual               58    23.95

TABLE 3.
One-way ANOVA of arcsine-transformed M. edulis Glu5' allele frequency
at 3 sites in Cobscook Bay across sampling dates in 2000.

Source     df    SS     [F.sub.ratio]      P

Month       3   0.655      41.00        0.0002
Residual    6   0.032

TABLE 4.
Two-way ANOVA of arcsine-transformed M. edulis Glu5' allele
frequency at 3 sites in Cobscook Bay in September, 2000.

                                                   Observed
   Source      df    SS    [F.sub.ratio]     P     power *

Site            2   0.13       2.66        0.108     0.44
Depth           2   0.01       0.25        0.785     0.08
Site * Depth    4   0.09       0.91        0.488     0.21
Residual       13   0.31

* [alpha] set to 0.05.

TABLE 5.
The number of mussel spat observed on top, middle, and bottom rope
segments at Shackford Head, Cobscook flay across 4 collection dates
in 2000. For comparison to the data or Kenchington et al. (2002)
the numbers are expressed on a per meter basis. [N.sub.total],
represents the number of all mussel spat per meter of rope.
The number of M. edulis ([N.sub.M. edulis]) versus M. trossulus
and hybrid spat ([N.sub.other]) was estimated using the frequency
of individuals with M. edulis di-locus genotypes at the Glu5'
and ITS markers.

Month       Depth    [N.sub.total]   % M. edulis

August      Top             5             --
            Middle         10            40.1
            Bottom          3             --

September   Top           638            10.3
            Middle        560            23.9
            Bottom        360            25.2

October     Top           143            90.4
            Middle        129            79.7
            Bottom         66            74.5
December    Top            16            95.0
            Middle         35            89.6
            Bottom        118            74.2

Month       Depth    [N.sub.M. edulis]   [N.sub.other]

August      Top
            Middle            4                 6
            Bottom

September   Top              64               574
            Middle          129               431
            Bottom           65               195

October     Top             129                14
            Middle          103                26
            Bottom           50                17

December    Top              15                 1
            Middle           32                 4
            Bottom           97                31

ACKNOWLEDGMENTS

This research was supported by the Maine Sea Grant program under contract 99-343 and Hatch funds provided by the US Department of Agriculture and the Maine Agriculture and Forestry Experiment Station (Project No. ME08510). The authors thank Tom Pottle, Heritage Salmon Inc. and Frank Ayres of Maine Salmon Inc. for access to their sites in Cobscook Bay and two anonymous reviewers for their helpful comments on a draft of this manuscript. Afton McGowen was supported by the Maine Research Internships for Teachers and Students Program (MERITS) sponsored by the Foundation for Blood Research. This is Maine Agricultural and Forest Experiment Station external publication #2724.

LITERATURE CITED

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

Boecklen, W. J. & D. J. Howard. 1997. Genetic analysis of hybrid zones: numbers of markers and power of resolution. Ecology 78:2611-2616.

Brooks, D. A., M. W. Baca & Y.-T. Lo. 1999. Tidal circulation and residence time in a macrotidal estuary: Cobscook Bay. Maine. Estuar. Coastal Shelf Sci. 49:647-665.

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.

Freeman. K. R., K. L. Perry, T. G. DiBacco & D. J. Scarratt. 1994. Observations on two mytilid species from a Nova Scotian mussel farm. Can. Tech. Rep. Fish. Aquat. Sci. 47.

Freeman, K. R. 1996. An examination of biological and other factors affecting mussel aquaculture development in the Scotia-Fundy region of Nova Scotia. Can. Tech. Rep.Fish. Aquat. Sci. 32.

Freeman, K. R., E. R. Kenchington & S. P. Macquarrie. 2002. Comparative settlement depths of Mytilus edulis C. Linnaens, 1758 and M. trossulus Gould, 1850: I. a mesocosm study. J. Shellfish Res. 21:59-65.

Heath, D. D., P. D. Rawson & T. J. Hilbish. 1995. PCR-based nuclear markers identify alein blue mussel (Mytilus spp.) genotypes on the west coast of Canada. Canadian Journal of Fisheries and Aquatic Science 52:2621-2627.

Hickman, R. W. 1992. Mussel Cultivation. In: E. M. Gosling, ed. The mussel Mytilus: ecology, physiology, genetics, and culture. Amsterdam: Elsevier Science Publishers. pp. 465-510.

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

Kenchington, E. R., K. R. Freeman. B. Vercaemer & B. MacDonald. 2002. Comparative Settlement Depths of Mytilus edulis C. Linnaeus, 1758 and M. trossulus Gould 1850: II. Field observations. J. Shellfish Res. 21:67-73.

Koehn, R. K., R. Milkman & J. B. Mitton. 1976. Population genetics of marine pelecypods. IV. Selection, migration and genetic differentiation in the blue mussel Mytilus edulis. Evolution 30:2-32.

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

Lutz, R. A. 1980. Mussel culture and harvest: a North American perspective. Amsterdam: Elsevier Science Publishers.

Mallet, A. L. & C. E. Carver. 1995. Comparative growth and survival patterns of Mytilus trossulus and Mytilus edulis in Atlantic Canada. Canadian Journal of Fisheries and Aquatic Sciences 52:1873-1880.

Mallet, A. L. & C. E. Carver. 1999. Maritime distribution and commercial production performance of Mytilus edulis and Mytilus trossulus. Bulletin of the Aquaculture Association of Canada 99:7-13.

Maloy, A. P., B. J. Barber & P. D. Rawson. 2003. Gametogenesis in a sympatric population of blue mussels, Mytilus edulis and Mytilus trossulus, from Cobscook Bay (USA). J Shellfish Res. 22:119-123.

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.

National Marine Fisheries Service. 2003. Fisheries statistics and economics division, annual commercial landings statistics. Available at http:// www.st.nmfs.gov/st1/commercial/landings/annual_landings.html

Penney, R. W. & M. J. Hart. 1999. Distribution, genetic structure, and morphomettry of Mytilus edulis and M. trossulus within a mixed species zone. J Shellfish Res. 18:367-384.

Penney, R. W., M. J. Hart & N. Templeman. 2002. Comparative growth of cultured blue mussels, Mytilus edulis, M. trossulus and their hybrids, in naturally occurring mixed-species stocks. Aquaculture Research 33: 693-702.

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

Rawson, P. D., S. Hayhurst & B. Vanscoyoc. 2001. Species composition of blue mussel populations in the Northeastern Gulf of Maine. J. Shellfish Res. 20:31-38.

Roberts, C. & C. Newell. 1999. The Maine guide to mussel raft culture. Rockland, Maine: Island Institute. 32 pp.

Sponaugle, S., R. K. Cowen, A. Shanks. S. G. Morgan, J. M. Leis, J. Pineda, G. W. Boehlert, M. J. Kingsford, K. C. Lindeman, C. Grimes & J. L. Munro. 2002. Predicting self-recruitment in marine populations: Biophysical correlates and mechanisms. Bull. Mar. Sci. 70(S):341-375.

Toro, J. E., R. J. Thompson & D. J. Innes. 2002. Reproductive isolation and reproductive output in two sympatric mussel species (Mytilus edulis, M. trossulus) and their hybrids from Newfoundland. Mar. Biol. 141: 897-909.

PAUL D. RAWSON, * AFTON MCGOWEN, AND CHRIS BARTLETT 

School of Marine Sciences University of Maine, Orono, Maine, 04469 and Maine Sea Grant Program Eastport, Maine 04631

* Corresponding author. E-mail: [e-mail]