Impact of green crab (Carcinus maenas L.) predation on a population of soft-shell clams (Mya arenaria L.) in the Southern Gulf of St. Lawrence.

ABSTRACT A caging experiment was carried out on an estuarine mudflat in Pomquet Harbour, Nova Scotia, Canada from late Mayto early September 2001. Six replicate 0.83 [m.sup.2] cages were set up for each of the following treatments: low predator density cages withone green crab (Carcinus maenas), high predator density with five green crabs, exclosure cages with no crabs, and control cages with sections of the sides removed. In addition, six 0.83 [m.sup.2] sections of exposed mudflat served as further controls. Green crabs significantly reduced numbers of small (<17 mm) soft-shell clams (Mya arenaria), removing ca. 80% of the small clams within low and high predator cages. There was no significant difference in large (>17 mm) soft-shell clam densities for any of the treatments. Green crabs consumed small clams at a minimum rate of 3.1 clams/crab/day and a maximum rate of 21.8/clams/crab/day. Predation intensity did not vary with density of crabs within cages.

KEY WORDS: green crab, Carcinus maenas, invasive, soft-shell clam, Mya arenaria, predation, caging experiment

INTRODUCTION

The establishment of an invasive or non-indigenous species can result in significant ecological changes in the receiving environment. The green crab, Carcinus maenas (Linnaeus), is an example of a very successful marine invasive species. After establishment in a new environment, green crabs have been shown to effect changes at the individual, population and community levels (Hughes & Elner 1979, Grosholz et al. 2000, Trussell & Smith 2000, McDonald et al. 2001).

Although green crabs directly and indirectly affect many benthic organisms, bivalves are usually the preferred prey (Ropes 1968, Elner 1981, Grosholz & Ruiz 1995) and exhibit dramatic post-invasion changes. Grosholz el al. (2000) found that invasive green crabs exert predatory control on the native clams Nutricola tantilla (Gould) and Nutricola confusa (Gray), population levels of these clams declined drastically within 3 years of the arrival of green crabs, and have shown no sign of recovery since the invasion. In Tasmania, Walton et al. (2002) used both field observations and manipulative field experiments to investigate green crab predation on the venerid clam Katelysia scalarina (Lamarck). They found that green crabs exerted considerable predatory pressure on smaller clams (<13 mm shell length). On the east coast of North America, declines in population sizes of the soft-shell clam (Mya arenaria Linnaeus) have been linked to the arrival of the green crab (Glude 1955).

It has been established that soft-shell clams are a significant prey item for green crab. Glude (1955) found excellent survival and growth in areas where green crabs were excluded by fencing or screening. Ropes (1968) examined approximately 4000 green crab stomachs, and concluded that soft-shell clams were an important component of the crab's diet in Massachusetts. He also reported that the frequency of M. arenaria remains in green crab stomachs was highest during September to November, and suggested that this was due to the abundance of juvenile clams in the fall. In a similar study from Port Hebert, Nova Scotia, Elner (1981) also found that soft-shell clams were an important prey item for green crabs. However, in contrast to the New England study, Elner (1981) reported significant predation on soft-shell clams in May and August.

In summary, correlations have been demonstrated between green crab arrival and declines in abundance of soft-shell clams. Furthermore, gut content studies have shown that soft-shell clams are a significant prey item for green crabs. We used an in situ caging approach to investigate the timing, nature, and extent of green crab predation on a population of soft-shell clams. Specifically, we wanted to: (1) experimentally determine the extent of green crab predation on soft-shell clams during the summer; (2) determine if green crab predation on soft shell clams was size selective; and (3) determine if there was a relationship between green crab density and predation rate.

METHODS

Pomquet Harbour, Nova Scotia, is a shallow estuary (tidal amplitude of approximately 1 m) that is connected by a narrow channel with St. Georges Bay (Fig. 1). The study site was a 7 x 11 m section of a lower intertidal flat of slightly gravelly sand located near the mouth of the Pomquet River (45[degrees]36.19'N, and 61[degrees]36.19'W). The surrounding terrestrial vegetation was predominately Spartina alterniflora (Loisel) with patches of Solidago sempervirens (L.), Convolvulus arvensis (L), Festuca pratensis (Hudson), and Rosa virginiana (Mill.). Salinity and temperature at the site were measured from high tide to low tide for one spring tide on August 5, 2001. The surface salinity ranged from 26.0 [per thousand] to 23.2 [per thousand] and the bottom salinity ranged from 29.0 [per thousand] to 25.1 [per thousand] during the ebb. The surface temperature ranged from 21.2[degrees]C to 26.3[degrees]C and the bottom temperature ranged from 21.8[degrees]C to 26.3[degrees]C. The water depth in the middle of the study site was approximately 35 cm at high tide, and the mudflat was approximately 20 cm above sea level at low tide.

[FIGURE 1 OMITTED]

On May 22, 2001, limited preliminary samples were collected to identify the macrofauna, with emphasis on M. arenaria. The sampling was limited partly to minimize disturbance on the relatively small clam flat, and partly because the design was not based on comparisons of pre- and post-experimental conditions. Six core samples (10 cm diameter, 20 cm depth) were taken from randomly selected areas around the proposed location for the cage matrix. The 6 core samples were washed through a series of graded sieves to a bottom mesh size of 0.5 mm, and all retained organisms were removed and preserved. This mesh size has been shown to yield close to 100% of the macrofaunal species and over 95% of the biomass (Reish 1959). Also, sediment from a 1 [m.sup.2] section of the site was removed to a depth of 20 cm and passed through a 1.27 cm sieve to collect the large M. arenaria (clams >17 mm). We designated clams larger than 17 mm as large soft-shell clams for the purpose of this experiment. This size represents the maximum observed size of that year's set of small, recently settled clams.

Cages were 0.91 x 0.91 x 0.30 m high, and were constructed of plastic-coated wire, with a square mesh opening of 1 x 1 cm. Treatments were as follows: 6 exclosure cages (E), with no predators added, 6 control cages (C1) with 20 x 70 cm portions of each side removed allowing for unrestricted predator movement, 6 undisturbed mudflat controls that were outside but adjacent to the cage matrix (C2), 6 low-predator-density cages (D1) with 1 crab added, and 6 high-predator-density cages (D2) with 5 crabs added. The low predator densities of 1.2 crabs /[m.sup.2] reflects mark and recapture data acquired the previous year in a similar habitat in an adjacent harbor (Campbell 2000). The high predator density of 6.1 crabs/[m.sup.2] reflects green crab densities estimated by Young et al. (1999) for a New England salt marsh tidal creek.

To reduce the potential impact of any environmental heterogeneity of the mudflat, a randomized block design was used for location of cages. During a spring low tide on May 23, 2001, the cages were deployed in a 6 x 4 rectangular matrix, with the four treatments randomly assigned to a position in each row (= block), and 1 m between all the cages. The rows were oriented perpendicular to the gradual slope of the mudflat, resulting in approximately 10 cm height difference between the highest and lowest cages. The cages were pushed into the sediment to a depth of 20 cm, allowing a 10-cm space above the substratum. Green crabs trapped the previous day were added to the appropriate cages by cutting a small flap in the cage, adding crabs, and closing the flap with plastic cable ties. Only male crabs (53.5 [+ or -] 5.22 (SD) mm, ranging from 44 mm to 65 mm in CW) with complete pairs of chelipeds and sets of walking legs were used in the enclosures. The crabs in the D1 and D2 cages were checked every spring tide throughout the next 3 months by gently probing the mud with a wire probe. Following Gee et al. (1985), any crab mortality during the experiment was noted, but crabs were not replaced.

Between August 21 to 23 (approximately 3 mo after deployment), the cages were removed from the substratum. To test for cage-induced physical changes to the site, 4 random sediment samples were taken (5 cm deep, 2.5 cm diameter) from each cage area, as well as the 6 C2 plots sampled outside of the cage matrix. These samples were frozen and later dried at 70[degrees]C for 7 hours. A weighed subsample was heated in a 500[degrees]C muffle furnace for 7 hours and re-weighed to determine the percent organic content, which was compared among treatments using the Kruskal-Wallis non-parametric H-test. Following Folk (1974), subsamples of the remaining sediment were then dry sieved through a series of Wentworth graded sieves to determine the size distribution of the particles. These size distributions were then compared with check for cage effects.

Three large cores (10 cm diameter, 20 tin depth) were taken from randomly selected locations within each cage (Hall et al. 1990). To avoid edge effects, cores were never taken within 15 cm of the sides of the cages (Kent & Day 1983). The core samples were sieved to a bottom mesh size of 0.5 mm, and clams preserved and counted. After coring, all sediment inside the cages was excavated to a depth of 30 cm and sieved through a 1.27-cm coarse sieve to collect large soft-shell clams. Large clams were counted and a concentric ring analysis was used to determine age (Newcombe 1936). In addition to sampling of cage sites, the six C2 plots, which were chosen randomly on all sides of the cage matrix, were subjected to the same sampling regimen.

Core sample data was used to determine densities of small soft-shell clams (clams <17 mm in length). Since three cores were taken from each cage, the arithmetic means of the cores were used in the analysis to avoid problems of pseudoreplication (Gee et al. 1985). The Shapiro-Wilk method was used to test for the normality of the count data for both size classes of clams. Data for large and small soft-shell clams was normally distributed (P = 0.13 and 0.42 respectively). An ANOVA appropriate for a randomized block design and the Tukey multiple comparison test was used to test for significant difference among treatments. Initially we only included the 4 cage treatments E, D1, D2, and C2 in the analysis. Because no block effect was detected in the initial analysis (P = 0. 157), a subsequent analysis was run with the additional data from the mudflat C2.

On May 30, 2002, the study site was again sampled to determine the fall/winter survival of the 2001 summer set of small soft-shell clams. Six random core samples were obtained adjacent to the study site and clams were removed, measured, and counted.

RESULTS

No noticeable scouring or deposition of sediment from water movement was observed in cages throughout the experiment. Sediment organic content averaged 3.30% [+ or -] 0.57 (SD), with no significant difference among treatments ([chi] = 2.612, d.f. = 5, P = 0.625). Based on particle size distribution, the sediment for the study site can be classified (Folk 1974) as slightly gravelly sand. Particle distribution analysis showed no difference between treatments (Fig. 2), No long-term fouling by growing or drift algae occurred on the cages.

[FIGURE 2 OMITTED]

The core samples taken before the cages were deployed yielded 3 species of macrofauna: the bivalves, M. arenaria and Macoma balthica L., and the polychaete, Hediste diversicolor Muller. Soft-shell clams made up more than 95% of macrofaunal numbers and biomass. Large soft-shell clams had a pro-experiment density of 21/[m.sup.2]. There were no small clams found at this time.

No significant difference was found among treatments for large soft-shell clams, both when only cage data were compared (Fig. 3, P = 0.101) and when undisturbed mudflat C2 data were included (P = 0.1341. Large soft-shell clam densities ranged from 11 to 24 clams/[m.sup.2], and had a mean density of 19.2 [+ or -] 6.05 (SD) clams/[m.sup.2], (see Fig. 3). A significant difference was round among treatments for small soft shell clams, when the cage data were compared, and when the undisturbed mudflat C2 was included (Fig. 4, P < 0.001 for both comparisons). Specifically, D1 and D2 enclosures had significantly fewer small clams than all other treatments but did not differ from each other. Small soft-shell clam densities ranged from 305/[m.sup.2] (D2 enclosure) to 1712 clams/[m.sup.2] (exclosure).

[FIGURES 3-4 OMITTED]

Some crab mortality occurred in enclosure cages during the last 3 weeks of the experiment, thereby making calculation of consumption rates difficult. Average estimated consumption rates were calculated using the initial and final crab densities. Two of six crabs in the D1 cages died during the last month, yielding a final D1 density of 0.80 crabs/[m.sup.2]. In the D2 treatments, which started with 5 crabs in each cage, the number of surviving crabs were 0, 1, 2, 2, 3, and 3 crabs, yielding a final D2 density of 2.16 crabs/[m.sup.2].

Because no significant difference was detected among treatments for large soft-shell clams, the age analysis data were pooled to provide a picture of the age structure at the study site (Fig. 5). The 4-year and 5-year age classes dominated the sampled soft shell clams on the study site, with very few clams aged 1, 2, and 3 years in the population. The spring 2002 sampling yielded 148 small soft-shell clams/[m.sup.2]. This spring density was significantly less than the late summer density of approximately 1500/[m.sup.2], and represents nearly 90% fall/winter mortality.

[FIGURE 5 OMITTED]

DISCUSSION

Small Soft-shell Clams

We found clear evidence of significant green crab predation on soft-shell clams with small clam densities within enclosures significantly lower than in all other treatments. This concurs with a summary of green crab feeding studies provided by Cohen et al. (1995), in which they report that green crabs select soft-shell clams less than 20 mm in length. Welch (1968) found that green crabs regularly destroyed sets of small clams, and Beal et al. (2001) states that losses of recently settled soft-shell clams coincided with the appearance of predators, including green crabs, within experimental units. Glude (1955) found a daily consumption rate of 15 small clams/crab in the laboratory. Our ability to calculate daily consumption rates is somewhat hampered by the crab mortality that occurred in the last 3 weeks of the study, and our inability to determine when clams were consumed. However, we can calculate a range by assuming either that all clams were consumed prior to crab mortality, or that all clams were consumed by the remaining crabs after the mortality occurred. Daily consumption rates were between 14.5 and 21.8 small clams/crab in the one crab treatments and 3.1 and 8.3 small clams/crab in the five crab treatments. These ranges are the first in situ estimates of green crab consumption rates on small soft shell clams.

One of the objectives of this research is to assess the extent of summer predation on soft-shell clams. A comparison of clam densities among exclosure cages, control cages, and open mudflat samples provide estimates of overall predation, and also gives some clue as to the type of predator. The control cages were designed to exclude avian predators, but allow access by fish and invertebrates. The open mudflat sites were available to all predators. No statistically significant differences were detected between exclosure clam densities and densities from the two types of control samples. Even though our results demonstrate that green crabs are capable of harvesting a significant number of small soft-shell clams from the flats, this activity does not take place during the summer. Ropes (1968) found soft-shell clams occurred most frequently in green crab stomachs during September to November. As well, Glude (1955) found that strongly recruiting populations of small soft-shell clams had disappeared by the autumn where green crabs were present. These reports, as well as this study, are in contrast with Elner (1981), who found that M. arenaria were a large component of green crab diet in both May and August. The difference in the timing of crab predation on soft shell clams in Elner's (1981) work may be due to much cooler summer water temperatures along the southeastern coast of Nova Scotia, as compared with the two New England studies, and our estuarine site in Pomquet Harbour. On August 5, maximum bottom temperatures at our study site were 26[degrees]C, and could well have exceeded 30[degrees]C in August. In the late fall, when the water temperatures are lower, green crabs may increase foraging on the clam flats. At this time, the crabs would also benefit from the increased size of young-of-the-year clams. Extensive pitting attributed to green crab foraging has been observed on other Pomquet Harbour mudflats during the late autumn and early winter (P. J. Williams, unpublished data).

To obtain an estimate of small clam mortality during the fall/ winter period, we sampled the study site the following spring (2002) before the green crabs became active. We found that approximately 90% of the small soft-shell clams had been removed in the period from August 2001 to May 2002. Although we cannot partition the removal among migratory birds, fish, green crab, other invertebrates, or physical factors such as ice scour, the loss is comparable with the 80% mortality we observed within green crab enclosures.

There was no difference in small soft-shell clam abundance between the low density (1 crab) and the high density (5 crabs) crab enclosures. Crab mortality during the experiment may have compromised our ability to detect differences. In retrospect, an experiment with a shorter time frame would have been better to elucidate density-related predation rates. However, even green crab densities of 0.8/[m.sup.2] can result in effective removal of 80% of small clams.

Large Soft-shell Clams

In contrast with the small soft-shell clam results, we found negligible predation by green crab on large soft-shell clams. The apparent selection of small clams over large ones has been suggested for green crab feeding on soft-shell clams (Welch 1968), and indeed fits a general pattern for crustaceans feeding on molluscs (Smith et al. 1999). There are a number of factors that might have led to the size discrimination. Green crabs are both tactile and chemosensory hunters (Cohen et al. 1995), and probably detect clams by following plumes from exhalent siphons, and/or coming in contact with siphons or siphon holes while probing sediment with appendages (Dare & Edwards 1981). During this study, the densities of small clams were much higher than the large clams, and therefore green crabs would have a higher encounter rate with the smaller clams. Studies that have addressed crab predation on molluscs from an energy optimization viewpoint (Elner & Hughes 1978. Juanes 1987) suggest that thicker shells in larger bivalves may lead to increased breaking time/energy expenditure, and may also result in chela damage for the crab. However, laboratory experiments have shown (Ropes 1968) that green crab are able to open and consume soft-shell clams that were longer than the carapace width of the crabs. In this study, the mean length of the large soft-shell clams was 54.7 [+ or -] 6.79 (SD) mm, and only 17 of the 463 clams sampled were larger than the 65 mm carapace width of the crabs used. Therefore, shell thickness was probably not a factor in green crabs selecting small clams. We feel that burial depth of the clams was the basis for the selection of small clams by green crabs. Zaklan and Ydenberg (1997) found similar results in predation experiments on soft-shell clams by red rock crabs (Cancer productus Randall) and attributed large clam survival to a depth refuge. Larger soft-shell clams are usually found deeper in the sediment than small clams (Blundon & Kennedy 1982, Zaklan & Ydenberg 1997), and we commonly found large clams at depths of 30 cm when sampling. Even though green crabs can burrow effectively, digging pits up to 15 cm deep (Ropes 1968, Ropes 1988, Lindsay & Savage 1978); our results suggest that green crabs are not effective predators at the depths where large soft-shell clams reside.

Large Soft-shell Clam Population Structure

The large soft-shell clam population at our experimental site was dominated by the 4-year and 5-year age classes, with weak year classes for clams aged 1, 2, and 3 years. Yearly variability in age class strength, attributed to changes in reproductive success, has been well documented in soft-shell clams (Kube 1996, MacKenzie & McLaughlin 2000, Strasser et al. 2001), but the weak year classes over the previous 3 years correspond with the period in which green crabs became abundant in Pomquet Harbour (P. J. Williams, unpublished data). An alternative explanation involves the recent outbreak of Hemic neoplasia that has been reported by McGladdery et al. (2001) in soft shell clams in Atlantic Canada, however incidence of this disease was highest in northwestern Prince Edward Island and relatively low in the areas closest to Pomquet Harbour. We believe that green crab predation in the years leading up to our study may have strongly influenced the population structure of soft-shell clams in Pomquet Harbour. Both Glude (1955) and Welch (1968) report that green crabs effectively removed young-of-the-year soft-shell clams from affected beds. Reproductive output peaks late in life for soft-shell clams (Brousseau 1978), with larger clams providing most of the reproductive effort. In Pomquet Harbour, when the older cohorts senesce or die, larval supply to the bed will drastically decline.

Cage Artifacts

In any study involving the placement of structures on a mudflat, there is the potential for a variety of confounding physical and biologic effects (Hulberg & Oliver 1980). In this study, there was no significant difference in organic content between treatments, no noticeable scouring or deposition of sediment within the cages, and no post-experiment differences in sediment particle size distribution. The cages experienced little fouling and drift algal accumulation was minimal. These results suggest that cage-induced physical artifacts were minimized in this experiment.

Refuge and feeding by small epibenthic predators within cages is another possible confounding factor, and these types of organisms could have caused stone of the mortality we observed in small soft-shell clams. We did not observe any such predators during biweekly checks or during the final sampling of the cages, however organisms such as small mummichogs (Fundulus heteroclitus Say), sand shrimp (Crangon septemspinosa Say), and small green and rock crab (Cancer irrorutus Say) could have entered the cages at high tide, and exited prior to low tide. There are, however, several reasons why we feel that any impact of such predators on the experimental results was negligible. Mummichogs are very stout fish, and the 1-cm mesh-size of the cages would preclude fish larger than approximately 50 mm in length. Smaller fish could have preyed upon newly settled soft-shell clams in cages, but this length is at the smaller range of fish reported to be effective predators on juvenile soft-shell clams (MacKenzie & McLaughlin 2000). We observed very few sand shrimp or rock crab at the site at any time. Finally, if predators such as these had removed small soft-shell clams, one would expect that they would remove similar numbers from all cages. If anything, the only bias one might expect would be for predators to avoid cages with green crabs, thereby making it more difficult to demonstrate a significant effect in green crab enclosure cages.

Clearly the confinement of the crabs in the D1 and D2 cages contributed to the crab mortality that occurred. The bulk (19 of 21) of the mortalities took place in the 20 days prior to the end of the experiment, from August 3 to August 23, 2001. Two crabs died in the D1 cages, in the absence of other crabs, suggesting the combination of starvation and the high daytime temperatures during this period could have contributed to the mortality. Cannibalism was likely a factor in the higher mortality observed in the D2 cages.

Aquaculture Implications

Our results suggest that soft-shell clam aquaculture operations in the southern Gulf of St. Lawrence should protect small soft-shell clams until they reach an appropriate size and depth refuge whereby they cannot be attacked by green crabs. If green crabs are allowed to prey freely on soft-shell clam beds, they will probably decimate small clam stocks and eventually cause steep population declines. Because adult soft-shell clams seem to have a depth refuge from predation, the effects of a decline of small clams may not be noticed until well after the onset of the green crab invasion, when the older, stronger year classes start dying off and are not replaced. Recreational or commercial harvesters only retain large clams, and may not immediately notice a reduction in numbers of small clams. Excluding green crabs by fencing is an effective mechanism for preservation of small soft-shell clams (Glade 1955, Beal & Kraus 2002). With the current expansion of green crabs into areas of heavy bivalve aquaculture, such a precaution may be needed to preserve harvested populations.

ACKNOWLEDGMENTS

The authors thank Ashley Bouchie, Sarah Fraser, Scan Mitchell, and Jack MacNeil for assistance with the field work and Peter Kaiser, Dr. Barry Taylor, Dr. Mike Melchin, and Dr. David Garbary for technical assistance. We also thank NSERC fur the operating grant used to fund this research.

LITERATURE CITED

Beak B. F., M. R. Parker & K. W. Vencile. 2001. Seasonal effects of intraspecific density and predator exclusion along a shore level gradient on survival and growth of smalls of the soft-shell clam. Mya arenaria L., in Maine. USA. J. Exp. Mar. Biol. Ecol 264:133-169.

Beal, B. F. & M. G. Kraus. 2002. Interactive effects of initial size, stocking density, and type of predator deterrent netting on survival and growth of cultured juveniles of the soft-shell clam, Mya arenaria L.. in eastern Maine. Aquaculture 208:81-111.

Blundon, J. A. & V. S. Kennedy. 1982. Refuges for infaunal bivalves from blue crab, Callinectes sapidus (Rathburn), predation in Chesapeake Bay.. J. Exp. Mar Biol. Ecol. 65:67-81.

Brosseau, D. J. 1978. Population dynamics of the soft-shell clam Mya arenaria. Mar. Bio. 50:63-71.

Campbell. E. 2000. The ecology of the green crab, Carcinus maenas, and its invasion in the Southern Gulf of St. Lawrence. B.Sc. Hons. Thesis, Saint Francis Xavier University. Antigonish, Nova Scotia, Canada.

Cohen, A. N., J. T. Carlton & M. C. Fountain. 1995. Introduction, dispersal and potential impacts of the green crab Carcinus maenas in San Francisco Bay, California. Mar. Biol. 122:225-237.

Dare. P. J. & D. B. Edwards. 1981. Underwater television observations on the intertidal movements of shore crabs, Carcinus maenas, across a mudflat. J. Mat. Biol. Ass. U.K 61:107-116.

Elner, R. W. 1981. Diet of green crab Carcinus maenas (L.) flora Port Hebert, southwestern Nova Scotia. J. Shellfish Res 1:89-94.

Elner, R. W. & R. N. Hughes. 1978. Energy maximization in the diet of the shore crab, Carcinus maenas. J. Anita. Ecol. 47:103-116.

Folk. R. L. 1974. Petrology of Sedimentary Rocks. Hemphill, Austin, Texas.

Gee. J. M., R. M. Warwick, J. T. Davey & C. L. George. 1985. Field experiments on the role of the epibenthic predators in determining prey densities in an estuarine mudflat. Estuar. Coast. Shelf Sci. 21:429-448.

Gillis, D. J., J. N. Macpherson & T. T. Rattray. 2000. The status of green crab (Carcinus maenas) in Prince Edward Island in 1999. PEI Department of Fisheries and Tourism Fisheries and Aquacutture Division, Report 225.

Glude, J. B. 1955. The effects of temperature and predators on the abundance of the soft- shell clam, Mya arenaria, in New England. Trans. Am. Fish. Soc. 84:13-26.

Grosholz, E. D., G. M. Ruiz, C. A. Dean, K. A. Shirley. J. L. Maron & P. G. Connors. 2000. The impacts of a nonindigenous marine predator in a California Bay. Ecology 81:1206-1224.

Grosholz, E. D. & G. M. Ruiz. 1995. Spread and potential impacts of the recently introduced European green crab. Carcinus maenas, in central California. Mar. Biol. 122:239-247.

Hall, S. J., D. Raffaelli, M. R. Robertson & D. J. Basford. 1990. The role of the predatory crab, Liocarcinus depurator, in a marine food web. J. Anim. Ecol 59:421-438.

Hughes, R. N. & R. W. Elner. 1979. Tactics of a predator, Carcinus maenas and morphological responses of the prey, Nucella lapillus. J. Anita. Ecol. 48:65-78.

Hulberg, L. W. & J. S. Oliver. 1980. Caging manipulations in marine soft bottom communities. Can. J. Fish. Aquat. Sci. 37:1130-1139.

Kent, A. C. & R. W. Day. 1983. Population dynamics of an infaunal polychaete: the effect of predators and an adult-recruit interaction. J. Exp. Mar. Biol. Ecol. 73:185-203.

Kube, J. 1996. Spatial and temporal variations in the population structure of the soft shell clam Mya arenaria in the Pomeranian Bay (southern Baltic Sea). J. Shellfish Res. 35:335-344.

Lindsay, J. A. & N. B. Savage. 1978. Northern New England's threatened soft shell clam populations. Environ. Manage. 2:443-452.

McDonald, P. S., G. C. Jensen & D. A. Armstrong. 2001. The competitive and predatory impacts of the nonindigenous crab Carcinus maenas (L.) on early benthic phase Dungeness crab Cancer magister Dana. J. Exp. Mar. Biol. Ecol. 258:39-54.

McGladdery, S. E., C. L. Reinisch, G. S. MacCallum, R. E. Stephens, C. L. Walker & J. T. Davidson. 2001. Haemic neoplasia in soft shell clams (Mya arenaria): recent outbreaks in Atlantic Canada and discovery of a p53 gene homologue associated with the condition. Bull. Aquacult. Assoc. Can. 101:19 26.

Newcombe, C. L. 1936. Validity of concentric tings of Mya arenaria L. for determining age. Nature. 137:191-192.

Reish, D. J. 1959. A discussion of the importance of screen size in washing quantitative marine bottom samples. Ecology 40:307-309.

Ropes, J. W. 1988. The food habits of five crab species at Pettaquamscutt River, Rhode Island. Fish. Bull. 87:197-204.

Ropes, J. W. 1968. The feeding habits of the green crab, Carcinus maenas. Fish. Bull. 67:183-203.

Smith, T. E., R. C. Ydenberg & R. W. Elner. 1999. Foraging behaviour of an excavating predator, the red rock crab (Cancer products Randall) on soft-shell clam (Mya arenaria L.). J. Exp. Mar. Biol. Ecol. 238: 185-197.

Strasser, M., A. Hertlein & K. Reise. 2001. Differential recruitment of bivalve species in the northern Wadden Sea after the severe winter of 1995/96 and of subsequent milder winters. Helgol. Mar. Res. 55:182-189.

Trussel, G. C. & L. D. Smith. 2000. Induced defenses in response to an invading crab predator: An explanation of historical and geographical phenotypic change. Proc. Natl. Acad. Sci. U. S. A. 97:2123-2127.

VanBlaricom, G. R. 1982. Experimental analyses of structural regulation in marine sand community exposed to oceanic swell. Ecol. Monogr 53: 283-305.

Walton, W. C., C. MacKinnon, L. F. Rodriguez, C. Proctor & G. M. Ruiz. 2002. Effect of an invasive crab upon a marine fishery: green crab, Carcinus maenas, predation upon a venerid clam, Katelysia scalarina, in Tasmania (Australia). J. Exp. Mar Biol. Ecol. 272:171-189.

Welch, W. R, 1968. Changes in abundance of the green crab, Carcinus maenas (L.) in relation to the recent temperature changes. Fish. Bull. 67:337-345.

Young, T., S. Komarow, L. Deegan & R. Garritt. 1999. Population size and summer home range of the green crab. Carcinus maenas, in salt marsh tidal creeks. Biol. Bull 197:297-299.

Zaklan, S. D. & R. Ydenberg. 1997. The body size-burial depth relation ship in the infaunal clam Mya arenaria. J. Exp. Mar. Biol. Ecol. 215: 1-17.

TREVOR FLOYD AND JIM WILLIAMS *

Department of Biology, St. Francis Xavier University, Antigonish, Nova Scotia, B2G 2W5, Canada

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