Genotype-dependent survival, growth, and production in cultured blue mussels, Mytilus SPP.: results of a reciprocal seed transfer experiment.

ABSTRACT A reciprocal mussel seed transfer experiment was conducted involving two farm sites (Charles Arm and Thwart island) whose seedstocks differed significantly in their respective compositions of Mytilus edulis L., M. trossulus Gould and hybrids. Electrophoretic variation at the mannose phosphate isomerase (Mpi) enzyme locus was used to classify the two species and their hybrids. Survival and growth in shell length, wet weight, shell weight and flesh weight were monitored over a 14-mo period. Both stock groups originating from Thwart Island seed (TI reared on its native site and the transferred TI-X stock) and the transferred stock group originating from Charles Arm seed (CA-X) showed no evidence of genotype-dependent variability in survival whereas, in the Charles Arm seed reared on its native site (CA), the proportion of M. edulis declined in relation to M. trossulus. There was no evidence of either a survival advantage or disadvantage in hybrids. After 14 too, both transferred stocks (CA-X and TI-X) had comparatively higher frequencies of M. edulis and hybrids and lower frequencies of M. trossulus than did their respective nontransferred stocks (CA and TI) indicating a significant genotype x site interaction effect on survival. Significant species, stock, stock x time, species x time and stock x species x time effects on shell length, wet weight, shell weight and flesh weight occurred. Intrastock growth rates of M. trossulus were lower than M. edulis and hybrids for shell length (3 of 4 stocks), and for the 3 weight variables, wet weight, shell weight and flesh weight (4 of 4 stocks). These results support the conclusion that growth is genotype-dependent in rope-cultured mussel populations and that these differences in weight growth favoring M. edulis are maintained when such stocks are transferred to other sites. Mussel biomass (kg [m.sup.-1]) was similar for the CA, CA-X and TI-X stock groups and lower for the TI stock group. Rates of production (wet weight) over the 14-too growout period were similar for the CA, CA-X and TI stock groups and significantly higher for the TI-X stock group. In both stock groups originating from Thwart Island seed (TI reared on its native site and the transferred TI-X stock), the intrastock rates of production of M. edulis and hybrids were significantly greater than sympatric M. trossulus. However, in the transferred Charles Arm stock (CA-X), the rates of production were similar in M. edulis and M. trossulus and significantly greater in hybrids. In the Charles Arm stock reared on its native site (CA), the rate of production of M. trossulus was significantly greater than either M. edulis or hybrids. We conclude M. edulis and hybrids have intrinsically greater rates of weight growth, but not necessarily length growth, than does M. trossulus. Neither M. edulis nor M. trossulus nor hybrids have consistently greater rates of survival or production in suspended rope culture. These results are discussed in the context of a directed seed-stocking program involving the transfer of M. edulis seedstocks to sites whose native stocks have a high M. trossulus component as a means to enhance commercial mussel production within the M. edulis-M, trossulus hybridization zone in Atlantic Canada.

KEY WORDS: Mytilus, genotype-dependent, reciprocal transfer, survival, growth, production

INTRODUCTION

Two mytilid mussel species, Mytilus edulis and M. trossulus, are endemic to northeastern North America. They are sympatric throughout a broad overlapping distributional zone extending from Newfoundland through the Gulf of St. Lawrence, along the Atlantic coast of Nova Scotia and into the northeastern United States (McDonald et al. 1991, Mallet & Carver 1992, Bates & Innes 1995, Penney & Hart 1999, Rawson et al. 2001). Localized mussel populations are typically composed of mixtures of the two species, often varying widely in relative frequency between adjacent sites in close proximity to each other (Mallet & Carver 1992, Bates & Innes 1995, Penney & Hart 1999). Indeed, nearby sites often display intersite genetic heterogeneity on a magnitude scale comparable to sites much farther apart (Penney & Hart 1999). The two species also readily hybridize (Saavedra et al. 1996, Comesana et al. 1999, Penney & Hart 2002a, Penney et al. 2002) forming a highly diverse, genetically heterogeneous hybridization zone.

Survival is genotype-dependent in rope-cultured populations on farm sites within the hybridization zone (Penney & Hart 2002a). Thus, temporal genetic heterogeneity is likely a significant contributing factor to the extensive genetic population structuring found throughout the hybridization zone (Penney & Hart 1999). Furthermore, growth in shell length, wet weight, shell weight and flesh weight are also genotype-dependent with growth rates for these weight variables being greater in M. edulis than M. trossulus, whereas growth rates in hybrids are typically intermediate between the two (Penney et al. 2002). Thus, genotype-dependent processes may be a significant contributory factor influencing production variability among mussel farms (Penney & Hart 2002a). Currently, many mussel farms within the hybrid zone, particularly sites in Newfoundland, use seed collected from their native sites with little regard or knowledge of their species composition. Because the relative species composition of the native seedstocks at these sites may differ dramatically (Penney & Hart 1999), this practice may contribute to the highly variable intersite growth and production indices throughout the industry. A recommendation has been made that growout of transplanted seed derived from unispecific or high-ratio M. edulis stocks should increase farm site production and reduce intersite variability in comparison with the use of seed collected on-site from indigenous mixed-species stocks (Penney et al. 2002). The greatest improvement in production would be anticipated at sites whose native seedstocks are primarily M. trossulus.

However, this recommendation has been advanced as a result of work carried out with indigenous seedstocks reared on their native sites. As Penney et al. (2002) pointed out, concern exists that our present state of knowledge of relationships among genetic heterogeneity, environmental variability and the intrinsic range of adaptive physiological flexibility in mussels (Hawkins & Bayne 1992) is inadequate to reliably predict the subsequent production performance of transplanted seedstocks. Yet, achieving predictable results from intersite seed transfers is key to successful implementation of a directed seed transplantation program by industry to increase commercial production. Implicit in the above recommendation are the assumptions that (a) mussel performance indices, which control production (e.g., survival and growth) are significantly controlled by stock genetics and thus are predictable when transferred among sites and (b) the range of natural intersite environmental variability over geographic scales wherein intersite stock transfers are practically feasible to industry does not exceed the intrinsic scope of adaptive flexibility within mussel populations.

In the various mussel species being commercially cultivated, production indices are affected by variable growth and survival, both of which are known to be significantly influenced by stock-related factors (i.e., genetic variability; Euentes et al. 1992, Stirling & Okumus 1995, Mallet & Carver 1995, Myrand & Gaudreault 1995 and site-related environmental variability; Dickie et al. 1984, Mallet & Carver 1989, Mallet et al. 1987). However, supporting evidence is somewhat equivocal concerning the comparative survival and growth of mussel stocks when transplanted versus when reared indigenously. There is some supporting evidence that inter-stock mortality differences are maintained when mussel stocks are transferred to new sites (Mallet et al. 1990, Myrand & Gaudreault 1995) suggesting that among-site survival variation is more influenced by stock genetics than environmental factors. In a series of experimental stock transfers in Nova Scotia, Mallet et al. (1990) noted survival rates in several stocks were unaffected by transfer leading to their conclusion that these stocks seemed insensitive to the environmental gradient. However, Dickie et al. (1984), Mallet et al. (1987) and Mallet & Carver (1989) concluded variation in mussel growth rates were mainly attributable to site-related environmental effects rather than stock (genetic) effects.

On a practical level, the mussel culture industry based in Prince Edward Island has successfully transplanted mussel seed from designated seed production sites to other growout sites over geographic scales of several kilometers. However, in Prince Edward Island, both the donor and recipient sites have indigenous mussel stocks that are virtually unispecific M. edulis (Mallet & Carver 1999). Similarly, for other mussel species and locations worldwide where commercial industry routinely transfers seedstocks successfully, (e.g., M. galloprovincialis) in Spain (Fuentes et al. 1994) and Perna canaliculatus in New Zealand (Alfaro & Jeffs 2003), these are areas with relatively low interstock genetic diversity. However, within the Atlantic Canadian hybrid zone, it has previously been observed that pure M. edulis stocks transferred from different locations can vary substantially in terms of tissue and shell growth, mortality and production (Mallet & Carver 1999). Past stock transfers involving more genetically heterogeneous seed populations have sometimes yielded unexpected growth and mortality patterns (Mallet & Carver 1989; Johannesson et al. 1990, Kautsky et al. 1990), which would be highly undesirable in a commercial context. Mallet and Carver (1999) have noted that transferred pure M. edulis stocks may not necessarily outperform a local mixed-species stock suggesting the possibility of significant localized adaptations in resident stocks. Interpretation of these sometimes conflicting results has been complicated by inadequate knowledge of stock genetic composition leading to concern regarding the predictability of subsequent survival and growth indices in transplanted seedstocks. This is especially so when the candidate seed stock for transplanting is genetically quite different from the indigenous stock (e.g., a unispecific M. edulis seedstock when transplanted to a predominately M. trossulus site). Because high-ratio M. trossulus populations are very common in most areas outside the general vicinity of Prince Edward Island, this concern is of significant importance to further industry development in the Atlantic Canadian hybrid zone.

In this study, we attempt to address this issue. We have conducted a reciprocal transfer experiment involving two genetically heterogeneous seedstocks, one a high ratio M. edulis stock and the other a relatively higher ratio M. trossulus stock from within the M. edulis-M, trossulus hybridization zone in Notre Dame Bay, Newfoundland. A suitable unispecific M. edulis stock was unavailable for intersite transfers within our geographic area under current seedstock transfer regulations, so we had to content ourselves with a known local high-ratio M. edulis stock. We compare the growth, survival and production patterns of each stock and of the M. edulis, M. trossulus, and hybrid constituents within each of the resulting four stocks (two reared indigenously, two reciprocally transferred) over a 14-mo period of rope culture at two farm sites. Specifically, we test the hypothesis that growth, survival and production rates of M. edulis and possibly hybrids as well are intrinsically higher than sympatric M. trossulus and that these intrastock differences are maintained when stocks are transplanted to other sites. The results are used to determine whether the substitution of indigenous M. trossulus seedstocks through intersite transfer of predominately M. edulis seedstock for growout can be supported as a viable approach to enhance commercial farm production.

METHODS

Previously, mussel populations at two farm sites in Notre Dame Bay (Charles Arm and Thwart Island), Newfoundland, were identified as predominately M. edulis and M. trossulus respectively (Penney & Hart 2002a). In August 2000, a reciprocal transfer experiment was set up involving seed collected on ropes suspended subtidally at both sites. Four seed groups were monitored for survival and growth over a 14-too period: Charles Arm seed kept at Charles Arm (CA), Charles Arm seed transferred to Thwart Island (CA-X), Thwart Island seed kept at Thwart Island (TI), and Thwart Island seed transferred to Charles Arm (TI-X). Seed mussels from each stock were mechanically de-clumped to ensure randomization and then sleeved in standard commercial plastic mesh using current industry husbandry practices for rope culture of blue mussels. Seed mussel sleeves from each donor stock were then either placed in the water at Charles Arm or transferred to Thwart Island. Each sleeve was 1 m in length and was vertically suspended at 0.5 m intervals from horizontal mainline ropes with flotation sufficient to ensure stable suspension at depths of approximately 4-6 m.

Three replicate sleeves were taken from each donor seed stock for laboratory analysis at the time of sleeving. On subsequent sample dates (November 2000, May 2001 and October 2001), three replicate sleeves were retrieved from each of the four stock groups on their respective farm sites. All the mussels were removed from measured sections of each sleeve and counted. On each sample date, one hundred individuals were randomly selected from each sleeve per stock group and measured for shell length, drained wet weight, shell weight and flesh weight (total of 300 from each site x group combination). Shell length was measured with digital callipers (Mitutoyo, Aurora, IL, USA). Wet weights were obtained after opening the shell valves and allowing the shell cavity to drain on paper towel. Hepatopancreas tissue was excised from each selected mussel, lyophilized and stored at 5[degrees]C for later allozyme analysis. The remaining soft tissues were then removed and shell weights were obtained after drying overnight. Flesh weights were calculated by difference between wet weight and dry shell weight.

For allozyme analysis of mannose phosphate isomerase (Mpi, EC 5.3.1.8), a small amount of freeze-dried material was ground to a fine powder with 0.5 M Tris HCL pH 8.0 buffer containing 20% glycerol and 0.2% NAD. Subsequent electrophoresis and staining on cellulose acetate plates followed the general procedure of Hebert and Beaton (1989) for the mannose phosphate isomerase locus, although a constant current of 2 mA per plate was used during the electrophoretic runs. Allele nomenclature is similar to that used by previous authors (Koehn et al. 1984, McDonald & Koehn 1988). Individual mussels were classified as M. edulis, M. trossulus or their hybrids based on their Mpi electrophoretic patterns according to previously published references for North American populations (Varvio et al. 1988, McDonald et al. 1991).

Survival over time was calculated as the difference in mean number of mussels [sleeve.sup.-1] on each sample date compared with the initial sample at the time of sleeving. In this context, changes in survival do not distinguish between changes caused by mortality versus dropoff from the culture gear. Rate of production was defined as the rate of change in wet weight between the initial and final sleeve weights where initial and final sleeve weights were calculated as the mean individual mussel wet weight multiplied by the number of mussels present. Rate of production was then calculated from the formula: Production = (final wet weight-initial wet weight)/initial wet weight. All subsequent statistical analyses were run with the SAS software system (SAS Institute Inc. 1988). Interstock and intergenotype differences in initial frequency distributions and survival were compared with the chi-square statistic. ANOVA were used to compare interstock and intergenotype differences in initial size of seed mussels and production over time. Tukey tests were applied to determine which comparisons were statistically significant. ANCOVA (homogeneity of slopes models) were used to compare interstock and intergenotype differences in growth.

RESULTS

Survival

In August 2000, mussel seed from the two donor sites, Charles Arm (CA) and Thwart Island (TI), were sleeved volumetrically after de-clumping to ensure randomization of all stock. The initial seed sleeving density was 2.441 [m.sup.-1] and 2.471 [m.sup.-1] for CA and TI seedstocks respectively. These differences were not significant (Student's t-test, P > 0.05). The frequency distribution of M. edulis, M. trossulus and their hybrids differed significantly in the two seedstocks at the time of sleeving. The initial CA seedstock was comprised of 76.2% M. edulis, 7.4% M. trossulus and 16.4% hybrids. In comparison, the initial TI seedstock was 54.7% M. edulis, 26.7% M. trossulus and 18.6% hybrids ([chi square], P < 0.001) (Fig. 1). Over the 14-too growout period, genotype-dependent survival patterns became evident in the CA stock but not in the CA-X, T1 or TI-X stocks (Fig. 2). In the CA stock, the frequencies of M. edulis and hybrids declined to 65.9% and 14.7% respectively, whereas the frequency of M. trossulus increased to 19.4% ([chi square], P < 0.001). After 14 mo, the frequencies of M. edulis, M. trossulus and their hybrids were 70.2%, 20.4% and 9.4% in the CA-X stock; 48.3%, 35.0% and 16.7% in the TI stock and 57.5%, 23.1%, and 19.4% in the TI-X stock respectively. However, these frequency changes of Mytilus genotypes over the 14-too growout period in the CA-X, TI and TI-X stocks were not significant compared with their respective initial Charles Arm or Thwart Island seedstock ([chi square], P > 0.05).

[FIGURES 1-2 OMITTED]

Mean survival over the 14-too growout period was significantly lower in the transferred Thwart Island stock (TI-X, 32.6%) than for either of the stocks derived from Charles Arm seed (nontransferred CA, 43.7%; transferred CA-X, 46.8%) or the nontransferred stock from Thwart island seed (TI, 48.1%) (Fig. 2). Interstock differences in mean survival were not significant among the CA, CA-X and TI stocks (ANOVA, P > 0.05). Comparison of the genotypic frequency distributions of the CA versus CA-X stocks and the TI versus TI-X stocks after 14 mo gave evidence of a significant site effect on the survival of the respective Mytilus genotypes. After 14 too, the genotypic frequency distributions of the CA and CA-X stocks had significantly diverged from each other ([chi square], P < 0.001) and the TI and TI-X stocks were also significantly different ([chi square], P < 0.006). In both cases, the transferred stocks (CA-X and TI-X) had comparatively higher frequencies of M. edulis and hybrids and lower frequencies of M. trossulus than did their respective nontransferred stocks (CA and TI). This provides evidence of a significant site x genotype interaction effect on survival among the three Mytilus genotypes.

[FIGURE 2 OMITTED]

Growth

When sleeved in August 2000, mean shell length, wet weight, shell weight and flesh weight of seedstocks from both donor sites differed significantly (Tukey, P < 0.05) (Table 1). For all four variables, CA seed mussels were larger than the TI mussels. Within the CA seedstock, there was no evidence of species-related differences in either of the four variables. However, within the TI seedstock, M. trossulus mussels were significantly larger than M. edulis and hybrids in terms of shell length, wet weight and flesh weight, but interspecies differences in shell weight were not significantly different.

Variability in all 4 growth variables (shell length, wet weight, shell weight and flesh weight) were all significantly related to species, stock and growout time (ANOVA, Table 2). The interaction effects of species x time and stock x time were also significant for all four variables, whereas the interaction effect of stock x species x time was significant for the three weight variables but not shell length. The interaction term, species x stock, was not significant for any of the four variables. The goodness of fit ([R.sup.2]) ranged from 0.66-0.72 for the four models.

Analysis of variance homogeneity of slopes models revealed that, over the 14-mo growout period, the two stocks raised at the Charles Arm site (CA and TI-X) consistently grew faster than did the two stocks reared at Thwart Island (TI and CA-X) (Fig. 3). At Charles Arm, the shell length slope for the transferred TI-X was significantly greater than the slope of the indigenous CA stock but slopes for all three weight variables were not significantly different between the two stocks. At Thwart Island, slopes for the transferred CA stock and the indigenous TI stock were not significantly different for shell length but, for all three weight variables, the respective slope for the CA-X stock was significantly greater than the TI stock.

[FIGURE 3 OMITTED]

Analysis of variance homogeneity of slopes models were subsequently used to elucidate the interaction effects involving species within stocks over the 14-mo growout period. For shell length, within-stock model slopes for M. trossulus were significantly lower than both M. edulis and hybrids for all but the CA-X stock (Fig. 4). A consistent within-stock interspecific pattern emerged for all three weight variables. In each case, within-stock model slopes for wet weight (Fig. 5), shell weight (Fig. 6) and flesh weight (Fig. 7), were significantly lower for M. trossulus than for M. edulis and hybrids. For all four variables, slopes of M. edulis and hybrids were not significantly different from each other.

[FIGURES 4-7 OMITTED]

Production

After 14 months, the mussel biomass (mean wet weight x number of survivors) of the CA CA-X and TI-X sleeves ranged from 4.26-4.49 kg [m.sup.-1] but were not significantly different (Table 3). Mean mussel biomass of the TI stock group was significantly lower than either of the other three stock groups. The rate of production, defined as the increase in mussel wet weight [m.sup.-1] of sleeve, was not significantly different among the CA, CA-X and TI stock groups but was significantly greater in the TI-X stock group (Table 3). Within both stock groups originating from the Thwart Island seedstock (TI and TI-X), M. edulis and hybrids had significantly greater production rates than M. trossulus. However, in the CA-X stock group, hybrids had a higher rate of production than either of the other two species. In the CA stock group, M. trossulus had a significantly higher rate of production than either M. edulis or hybrids. Thus, neither M. edulis nor M. trossulus nor their hybrids consistently exhibit greater rates of production within all stock groups.

DISCUSSION

Whether being considered from an individual farm or industry-wide perspective, the commercial industry objective is to maximize on-site production, while reducing production variability within and among farm sites. Genetic heterogeneity within and among sites can be a highly significant source of phenotypic variability in survival, growth and shell morphology (Penney & Hart 1999; Penney & Hart 2002a, Penney et al. 2002), all of which affect commercial industry production. At sites within mussel hybrid zones, genetic heterogeneity and phenotypic variability are both typically much greater than in areas occupied by single species (Gardner 1996). Because M. edulis exhibits faster growth in comparison with M. trossulus (Penney et al. 2002), it has been recommended that commercial production indices within the hybrid zone may be improved and intrasite and intersite variability reduced if industry were to convert to use of seed from unispecific or at least high-ratio M. edulis seedstock(s) replacing their present practice of seed collection from a multitude of sites, which often vary widely in species proportions (Penney et al. 2002). This would be especially beneficial at sites whose native mussel populations are predominately M. trossulus.

Commercial success of such a proposition depends on the relatively greater growth performance of M. edulis versus M. trossulus when stocks are grown on their native sites (Penney et al. 2002) being maintained when transplanted and for this growth differential to translate into increased production at the recipient site. The present work tests this proposition by assessing the relative growth, survival and production dynamics of sympatric cohorts of these two species and their hybrids in a reciprocal transplant experiment. With respect to survival, we confirm earlier observations (Mallet & Carver 1995; Penney & Hart 2002a) that genotype-dependent survival patterns occur among the various Mytilus genotypes in rope-cultured mussel populations within the Atlantic Canadian hybrid zone. However, these patterns do not consistently favor survival of either M. edulis or M. trossulus or their hybrids at all sites. In other words, survival is influenced by a significant stock x site interaction effect. It is also interesting to note that genotype-dependent selection favoring M. trossulus occurred in the Charles Arm stock reared on its native site. However, in a similar growout experiment involving an earlier year class of Charles Arm seed grown at the same site (Penney & Hart 2002a), there was no evidence of any selective survival pattern at the Mpi locus although overall survival of that Charles Arm year class was lower than the Thwart Island stock at that time. We infer from these results that genotype-dependent survival patterns are also temporally variable among year classes within the same stock, possibly indicative of a genotype-dependent interaction with changing environmental conditions. Thus, in a commercial context, year-class to year-class survival patterns within rope-cultured mussel populations are currently somewhat unpredictable. Because variable survival patterns may strongly affect production rates, this has important consequences for mussel aquaculture development within the hybrid zone.

The present results support earlier observations (Penney et al. 2002) that growth rates in wet weight, shell weight and flesh weight, but not necessarily growth in shell length, are typically greater in M. edulis and hybrids than in M. trossulus. We further conclude these genotype-dependent differences are maintained when stocks are transplanted to other sites. But, the intrinsically greater weight growth rates of M. edulis and hybrids do not necessarily translate into greater production performance in comparison with M. trossulus. Our analysis of production performance did not find either M. edulis or M. trossulus, or their hybrids consistently had comparatively higher rates of production in all cases. M. edulis and hybrid production rates were higher than M. trossulus in two of tour stock groups, M. edulis and M. trossulus production rates were similar in a third group, and the M. trossulus production rate was greater than either of M. edulis or hybrids in the fourth group. The lower rates of production of M. edulis within the two stock groups originating from the Charles Arm seedstock (CA and CA-X) and the relatively high production rate of M. trossulus within the CA stock group was largely because of the relatively lower survival of M. edulis within these two groups and to the relatively greater survival of M. trossulus in the CA stock group. The greater wet weight of the Charles Arm seed compared with Thwart Island seed at the time of sleeving also may have played a part in the comparatively lower production rates of their respective stock groups. Thus, we conclude our results do not support the proposition that M. edulis has intrinsically higher rates of production than M. trossulus. Therefore, a directed seed transfer program transplanting M. edulis stocks to sites whose native seedstocks are mostly M. trossulus will not necessarily result in increased rates of production on such mussel farms.

Lower survival rates of the M. edulis cohort compared with M. trossulus in mixed-species stocks has previously been reported from Lunenberg Bay in Nova Scotia (Mallet & Carver 1995). However, the relatively poor survival of M. edulis within stocks originating from Charles Arm seed is inexplicable at this point. This is especially so in light of the fact there was no evidence of genotype-dependent selection at the Mpi locus in a previous year class reared from seed to harvest size at Charles Arm using identical husbandry techniques and mussels of approximately the same age (Penney & Hart 2002a). M. edulis and M. trossulus are known to differ in their relative tolerance to temperature and salinity variations, with M. trossulus considered to be adapted to relatively colder, lower or more variable salinity environments (Vainola & Hvilson 1991, Mallet & Carver 1995, Rawson et al. 2001). However, it should be emphasized at this point that, although the transfer of stocks across large environmental gradients is known to have potentially significant deleterious effects on mussel survival (Theisen 1978, Kautsky et al. 1990, Johannesson et al. 1990), the two sites used in the present work are both located in the same bay and subjected to similar thermal and salinity regimes. Thus, no obvious strong environmental gradient exists between the Thwart Island and Charles Arm sites, which might offer a likely explanation for the results of the present study.

Genotype-dependent mortality has previously been reported in mixed M. edulis-M, trossulus stocks (Pedersen et al. 2000) and temporal and/or stock-related variability in its expression may be commonplace within Atlantic Canadian mussel stocks. This is evidenced by a reciprocal seed transfer experiment by Myrand and Gaudreault (1995) who found that four mussel stocks originating from nearby sites in the Magdalen Islands differed dramatically in resistance to summer mortality despite all being predominately M. edulis (Tremblay et al. 1998). In a subsequent transfer experiment, it was noted the resistant stock suffered a significant decrease in degree of heterozygosity when transferred to other nearby sites leading to the hypothesis that more heterozygous individuals migrate to the outer regions of sleeves faster than more homozygous individuals where they are more prone to drop-off caused by turbulence (Tremblay et al. 1998). However, Penney and Hart (2002a) evaluated multilocus heterozygosity at five of the same loci used by Tremblay et al. (1998) and found no evidence of temporal changes in heterozygosity in rope-cultured populations of either the Charles Arm or Thwart Island stocks. Thus, similar heterozygosity-fitness relationships in the Charles Arm stock appear an unlikely explanation for the relatively poor survival of the M. edulis component at Charles Arm.

Clearly, some factor other than simply genotypic variability at the Mpi locus results in significant intrastock and temporal survival variability in mixed M. edulis-M, trossulus stocks, although that unknown factor does seem linked to allelic variability at the Mpi locus. Allelic variability at several allozyme loci are known to be correlated with Mpi allelic variability in North American mussels (Varvio et al. 1988, McDonald et al. 1991) including native stocks at the two sites used in the present work (Penney & Hart 1999). Genotype-dependent selection at one such locus, Gpi, favoring survival of electrophoretically slower alleles has previously been noted in rope-cultured mussels from both Charles Arm and Thwart Island (Penney & Hart 2002a). Based on experimental work showing the enzyme products of different Gpi alleles had different thermal optima (Hoffman 1984, Hall 1985), Penney and Hart (2002a) advanced the hypothesis that suspended mussel culture may selectively favor individuals with electrophoretically slower Gpi alleles, whereas individuals with electrophoretically faster alleles would have an advantage in intertidal shore populations exposed to high air temperatures during emersion.

Electrophoretically slower Gpi alleles are associated more with M. trossulus than with M. edulis (Varvio et al. 1988, McDonald et al. 1991, Penney & Hart 1999), and the incidence of electrophoretically faster alleles is higher in the native Charles Arm stock than in Thwart Island stock. This could offer a potential physiological explanation for the relatively lower survival of the stocks originating from Charles Arm seed. While purely speculative, the hypothesis that slower Gpi electromorphs may be favored in subtidal conditions is indirectly supported by comparison of Mytilus Gpi electromorphs with those of the horse mussel, Modiolus modiolus. The natural spatial distribution of the latter is subtidal (Minor 1950) and hence subjected to a colder, relatively narrow range of water temperatures than the typically intertidal distribution of North American mytilid populations. Whether coincidence or not, all 11 electromorphs identified at the Gpi locus in Newfoundland populations of M. modiolus are electrophoretically slower than those of neighboring mytilids (Penney & Hart 2002b).

In summary, the present work supports the conclusion that the relatively faster rates of weight growth in rope-cultured M. edulis and hybrids compared with M. trossulus exhibited on their native sites are maintained when transferred to other sites, at least insofar as the donor and recipient sites are not substantially different in environmental conditions. However, the relative survival of the two species and hybrids and hence their relative production performance do not appear predictable based on their Mpi genotype alone. This conclusion is consistent with that of Mallet and Carver (1999) who observed that several pure M. edulis stocks from Prince Edward Island, New Brunswick and Nova Scotia when transferred to a site whose native population was primarily M. trossulus did not consistently outperform the local mixed-species stock. Therefore, selection of high performance seedstocks appropriate for widespread use by industry remains problematic and a need exists to conduct test transfers of each stock individually or, alternatively, greater understanding of the causal mechanisms of variable survival in mixed-species populations is required. However, because the greater shell weight of M. edulis may translate into lower levels of shell breakage during processing (Mallet & Carver 1995), selection of M. edulis stocks over local M. trossulus or mixed-species stocks for commercial growout appears likely to be a better alternative.

ACKNOWLEDGMENTS

The authors thank the owners and staff of Thimble Bay Farms Ltd. and Noel Bros. Mussel Farms Ltd. for their support and assistance maintaining experimental mussel gear at their respective farm sites.

LITERATURE CITED

Alfaro, A. C. & A. G. Jeffs. 2003. Variability in mussel settlement on suspended ropes placed at Ahipara Bay. Northland, New Zealand. Aquaculture 216:115-126.

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

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.

Dickie, L. M., P. R. Boudreau & K. R. Freeman. 1984. Influence of stock and site on growth and mortality in the blue mussel (Mytilus edulis). Can. J. Fish. Aquat. Sci. 41:134-140.

Fuentes, J., I. Reyero, C. Zapata & G. Alvarez. 1992. Influence of stock and culture site on growth rate and mortality of mussels (Mytilus galloprovincialis Lmk.) in Galicia, Spain. Aquaculture 105:131-142.

Fuentes, J., I. Reyero, C. Zapata & G. Alvarez. 1994. Production traits of the mussel Mytilus galloprovincialis cultured in Galicia (NW of Spain): relative effects of source of seed and growing environment. Aquaculture 122:19-31.

Gardner, J. P. A. 1996. The Mytilus edulis species complex in southwest England: effects of hybridization and introgression upon interlocus associations and morphometric variation. Mar. Biol. 125:385-399.

Hall, J. G. 1985. Temperature-related kinetic differentiation of glucose-phosphate isomerase alleloenzymes isolated from the blue mussel. Mytilus edulis. Biochem. Genet. 23:705-728.

Hawkins, A. J. S. & B. L. Bayne. 1992. Physiological processes and the regulation of production. In: E. Gosling, editor. The mussel Mytilus: ecology, physiology, genetics, and culture. Amsterdam: Elsevier Science. pp. 171-222.

Hebert, P. D. N. & M. J. Beaton. 1989. Methodologies for allozyme analysis using cellulose acetate electrophoresis. Beaumont, Texas: Helena Laboratories. 32 pp.

Hoffman, R. J. 1984. Thermal adaptation and the properties of phosphoglucose isomerase allozymes from a sea anemone. In: P. E. Gibbs, editor. Proceedings of the 19th European Marine Symposium. Plymouth, Devon, U. K. September 16-21, 1984. pp. 505-514.

Johannesson, K., N. Kautsky & M. Tedengren. 1990. Genotypic and phenotypic differences between Baltic and North Sea populations of Mytilus edulis evaluated through reciprocal transplantations. II. Genetic variation. Mar. Ecol. Prog. Ser. 59:211-219.

Kautsky, N., K. Johannesson & M. Tedengren. 1990. Genotypic and phenotypic differences between Baltic and North Sea populations of Mytilus edulis evaluated through reciprocal transplantations. I. Growth and morphology. Mar Ecol. Prog. Ser. 59:203-210. Koehn, R. K., J. G. Hall, J. Innes & A. J. Zeta. 1984. Genetic differentiation of Mytilus edulis in eastern North America. Mar. Biol. 79:117-126.

Mallet. A. L. & C. E. A. Carver. 1989. Growth, mortality, and secondary production in natural populations of the blue mussel, Mytilus edulis. Can. J. Fish. Aquat. Sci. 46:1154-1159.

Mallet. A. L. & C. E. A. Carver. 1992. The ecological and commercial implications of genetically distinct mytilids: completion of the geographical survey. Department of Fisheries and Oceans Aquaculture Resource development Branch, Biological Sciences Directorate, Ottawa, Ontario, Canada. 53 pp.

Mallet, A. L. & C. E. 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. Bulletin off the Aquaculture Association of Canada 99-3:7-13.

Mallet, A. L., C. E. A. Carver, S. S. Coffen & K. R. Freeman. 1987. Mortality variations in natural populations of the blue mussel, Mytilus edulis. Can. ,L Fish. Aquat. Sci. 44:1589-1594.

Mallet, A. L.. C. E. A. Carver & K. R. Freeman. 1990. Summer mortality of the blue mussel in eastern Canada: spatial, temporal, stock, and age variation. Mar. Ecol. Prog. Ser. 67:35-41.

McDonald, J. H. & R. K. Koehn. 1988. The mussels Mytilus trossulus and M. galloprovincialis 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 the three species of Mytilus in the northern and southern hemispheres. Mar. Biol. 111:323-333.

Minor, R. W. 1950. Field book of seashore life. G. P. Putnam and Sons. New York: Van Rees Press. 888 pp.

Myrand. B. & J. Gaudreault. 1995. Summer mortality of blue mussels (Mytilus edulis Linneaus. 1758) in the Magdalen Islands (southern Gulf of St. Lawrence. Canada). J. Shellfish Res. 14:395-404.

Pedersen, E. M., H. L. Hunt & R. E. Scheibling. 2000. Temporal genetic heterogeneity within a developing mussel (Mytilus trossulus and M. edulis) assemblage. J. Mar. Biolog. Assoc. U.K. 80:843-854.

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.

Penney, R. W. & M. J. Hart. 2002a. Is survival genotype-dependent in North American populations of farmed blue mussels, Mytilus spp.? J. Shellfish Res. 21:41-49.

Penney, R. W. & M. J. Hart. 2002b. Genetic variability in Modiolus modiolus populations from Newfoundland (Canada) revealed by protein electrophoresis with comparative notes on Mytilus spp. Canadian Technical Report of Fisheries and Aquatic Sciences 2430. 21 pp.

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. Aquac. Res. 33:693-702.

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.

Saavedra, C., D. T. Stewart, R. R. Stanwood & E. Zouros. 1996. Species-specific segregation of gender-associated mitochondrial DNA types in an area where two mussel species (Mytilus edulis and M. trossulus) hybridize. Genetics 143:1359-1367.

SAS Institute Inc. 1988. SAS/STAT user's guide, release ed. 6.03. Cary, North Carolina: SAS Institute Inc. 1028 pp.

Stirling, H. P. & I. Okumus. 1995. Growth and production of mussels (Mytilus edulis L.) suspended at two salmon cages and shellfish farms in two Scottish lochs. Aquaculture 134:193-210.

Theisen, B. F. 1978. Allozyme clines and evidence of strong selection in three loci in Mytilus edulis L. (Bivalvia) from Danish waters. Ophelia 17:135-142.

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

Vainola, R. & M. M. Hvilsom. 1991. Genetic divergence and a hybrid zone between Baltic and North Sea Mytilus populations (Mytilidae: Mollusca). Biol. J. Linn. Soc. Lond. 43:127-148.

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

R.W. PENNEY,* M.J. HART AND N.D. TEMPLEMAN

Department of Fisheries and Oceans, Science Branch, P. O. Box 5667, St. John's, Newfoundland, Canada, A1C 5X1

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

TABLE 1.
Initial size comparison of mussel seed from the two donor sites
(Charles Arm, CA and Thwart Island, TI) at the time of sleeving:
(a) full stock basis, and (b) and (c) by species within site. Numbers
with different superscripts are significantly different within columns
for each of (a), (b) and (c), P < 0.05 (Tukey multiple range test).

                                    Variable

                    Shell       Wet       Shell      Flesh
Stock              Length      Weight     Weight     Weight

(a)
  CA              26.42 (a)   1.28 (a)   0.45 (a)   0.82 (a)
  TI              21.53 (b)   0.81 (b)   0.32 (b)   0.50 (b)
(b) CA
  M. edulis       26.29 (a)   1.27 (a)   0.45 (a)   0.82 (a)
  Hybrid          26.30 (a)   1.24 (a)   0.44 (a)   0.80 (a)
  M. trossulus    28.37 (a)   1.48 (a)   0.52 (a)   0.96 (a)
(c) TI
  M. edulis       21.07 (a)   0.76 (a)   0.31 (a)   0.45 (a)
  Hybrid          20.53 (a)   0.67 (a)   0.27 (a)   0.41 (a)
  M. trossulus    23.49 (b)   1.05 (b)   0.37 (a)   0.68 (b)

TABLE 2.
Analysis of variance (ANOVA) test results for species
(M. edulis, hybrid, and M. trossulus) and stock (CA,
CA-X, TI, and TI-X) variation over the 14-month growout
period for shell length, wet weight, shell weight, and
flesh weight.

                                  F Value

                                                     Sp. x Time
                                                      ([double
Variable     Species        Stock         Time        dagger])

Shellen      7.76 **      60.58 ***    5403.99 ***    25.81 ***
Wetwgt       4.79 **      15.42 ***    5257.92 ***    64.28 ***
Shellwgt     3.24 *       11.08 ***    5479.08 ***    93.39 ***
Fleshwgt     5.65 **      17.71 ***    4470.41 ***    38.21 ***

                                  F Value

                                         Stk. X
           Sp. x Stk.    Stk. x Time   Sp. x Time
            ([double      ([double      ([double
Variable    dagger])      dagger])      dagger])      [R.sup.2]

Shellen       1.32        19.92 ***      1.99           0.69
Wetwgt        0.96        75.71 ***      5.68 ***       0.70
Shellwgt      0.74        96.29 ***      8.46 ***       0.72
Fleshwgt      1.17        53.53 ***      3.45 **        0.66

Numbers without superscript are not significantly different,
P > 0.05; *** P < 0.0001; ** P < 0.01; * P < 0.05. ([double
dagger]) sp. = species, stk = stock, etc.

TABLE 3.
Initial sleeve weight (mean wet weight x [number.sup.-1] of sleeve),
final mussel biomass after 14 months growout (mean wet weight x
[number.sup.-1] surviving), and rate of production ((final weight -
initial weight)/initial weight) of M. edulis, M. trossulus, hybrids,
and all species combined in each of transferred and non-transferred
stock groups. Numbers with letter and numeric superscripts are
significant for each variable across rows or down columns respectively
(Tukey, P < 0.05).

            Initial Weight (kg [m.sup.-1] sleeve)

Species         CA & CA-X      TI & TI-X

M. edulis          0.99           0.54
Hybrid             0.21           0.16
M. trossulus       0.11           0.37
All                1.31           1.07

                            Biomass (kg [m.sup.-1] sleeve)

Species        CA            CA-X            TI            TI-X

M. edulis     3.15           3.02           2.00           2.74
Hybrid        0.71           0.88           0.68           0.86
M. trossulus  0.63           0.36           1.09           0.71
All           4.49 (a)       4.26 (a)       3.77 (b)       4.31 (a)

                                   Rate of Production

Species            CA           CA-TX            TI            TI-X

M. edulis     2.18 (a) (1)   2.05 (a) (1)   2.70 (b) (1)   4.06 (c) (1)
Hybrid        2.41 (a) (1)   3.24 (b) (2)   3.21 (b) (2)   4.31 (c) (1)
M. trossulus  4.64 (a) (2)   2.19 (b) (1)   1.98 (b) (3)   0.93 (c) (2)
All           2.43 (a)       2.25 (a)       2.53 (a)       3.03 (b)