E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 498 2013 NORTHEASTERN NATURALIST 20(3):498–510 Influence of Sediment Type on Antipredator Response of the Softshell Clam, Mya arenaria Elsie Thomson1 and Damon P. Gannon1,* Abstract - Mya arenaria (Softshell Clam) inhabit a wide range of intertidal and subtidal sediment types in the western North Atlantic. They avoid predation by burrowing deeply in the sediment. We investigated the effect of sediment type on the antipredator responses of Softshell Clams to Carcinus maenas (Green Crab) as well as the relative costs and benefits of living in different sediment types. Clam burrowing depth, growth, and Green Crab predation rate were observed in experimental plots of mud, sand, and gravel. Clams exposed to crabs burrowed deeper than did control clams in all sediment types, but clams burrowed deepest in finer sediment types. Clams in coarser sediments also had thicker shells and suffered lower rates of predation than did those in mud. These results suggest distinctive costs and benefits associated with inhabiting different sediment types. For Softshell Clams, coarse sediments are most costly metabolically, but have lower predation risk compared to finer sediments. Introduction Sessile and slow-moving benthic animals use a variety of strategies to evade predation. For example, Mercenaria mercenaria Schumacher (Hard Clams) reduce feeding behavior to become cryptic when predators are perceived (Smee and Weissburg 2006) and increase shell thickness (Carriker 1959). Protothaca staminea Conrad (Pacific Littleneck) and Saxidomus giganteus Deshayes (Butter Clam) grow to reach size refuges where predators are unable to prey upon them (Boulding 1984). Mytilus edulis L. (Blue Mussel) invest in large adductor muscles to hold their shell tightly closed against predatory sea stars, which pry their valves open (Freeman 2007). However these strategies incur costs by either decreasing an organism’s ability to feed or increasing its energetic expenditures due to growth or locomotion (Harvell 1990, Sih 1987). Inducible antipredator responses allow organisms to minimize costs by selectively responding only to specific threats. These inducible responses can be morphological or behavioral and can take place over a range of time scales from milliseconds to weeks (Dill and Gillett 1991, Freeman 2007, Griffiths and Richardson 2006). Many marine species, including gastropods (Appleton and Palmer 1988) and bivalves (Flynn and Smee 2010, Freeman 2007, Griffiths and Richardson 2006), exhibit inducible responses to highly specific chemical cues produced by their predators. For example, Blue Mussels increased shell thickness in the presence of Carcinus maenas L. (Green Crab), which crush mussel shells (Freeman 2007). However, in the presence of Asterias vulgaris L. (Northern Sea Star), which pry mussels open, mussels increased the strength of their adductor muscles. 1Bowdoin Scientific Station and Department of Biology, Bowdoin College, 6500 College Station, Brunswick, ME 04011. *Corresponding author - dgannon@bowdoin.edu. 499 E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 Furthermore, studies have shown that bivalves respond specifically to predators that are feeding on conspecifics rather than simply to the presence of a predator (Griffiths and Richardson 2006). Mya arenaria L. (Softshell Clam) is a common bivalve species found along the western coast of the North Atlantic from South Carolina to the subarctic. They burrow in soft sediment throughout the intertidal and nearshore subtidal zones and are suspension feeders, using a siphon to filter algae from the water (Kamermans 1994). Softshell Clams burrow by extending a muscular foot between their gaping valves (Trueman 1954). Major predators of Softshell Clams include the snail Lunatia heros Say (Common Northern Moon Snail; Commito 1982), the nemertean worm Cerebratulus lacteus (Leidy) (Milky Ribbon Worm; Rowell and Woo 1990), seabirds, several species of crabs (Holland et al. 1980, Smith et al. 1999, Welch 1968), and humans. Green Crabs are invasive to the western North Atlantic and have been associated with the decline of Softshell Clam populations (Ropes 1968, Glude 1955). The distributions of Green Crabs and Softshell Clams overlap substantially in the western Atlantic. The Green Crab can be found from the mid-Atlantic states to Nova Scotia (Gosner 1987) and is a significant predator of Softshell Clams (Floyd and Williams 2004, Glude 1955, Ropes 1968). Unlike Pacific Littleneck and Butter Clams, which reach size refuges whereby the clam’s size prevents predators from eating them, large size alone does not appear to decrease predation risk for Softshell Clams (Blundon and Kennedy 1982b, Boulding 1984). The Softshell Clam’s brittle shell, which constantly gapes open, makes it vulnerable at all sizes. However, Softshell Clams can decrease their risk of predation from crabs by burrowing deep into the sediment (Flynn and Smee 2010, Whitlow et al. 2003, Zaklan and Ydenberg 1997). Blundon and Kennedy (1982b) found that Softshell Clams that were ≥10 cm below the sediment surface had significantly lower risk of predation by Callinectes sapidus M.J. Rathbun (Blue Crab) than did Softshell Clams at shallower depths. However, increasing burrowing depth decreases Softshell Clams’ feeding efficiency (Zaklan and Ydenberg 1997). Furthermore, many local habitat characteristics influence how Softshell Clams balance the trade-off between predation risk and feeding efficiency: e.g., salinity (Mathiessen 1960); season and temperature (Beal et al. 2001, Blundon and Kennedy 1982b, Brousseau 1979); dissolved oxygen concentration (Taylor and Eggleston 2000); presence of algal mats (Auffrey et al. 2004); and sediment type (Newell and Hidu 1982, Swan 1952). Sediment type affects Softshell Clam growth, movement, and vulnerability to predation. Softshell Clams grow slower in coarse sediments than they do in fine ones (Newell and Hidu 1982, Swan 1952). This slower growth may be due to increased costs of maintenance and locomotion. Coarser sediments cause chipping of the shell as well as resistance to shell gaping, which is required for extending the foot for burrowing (Swan 1952, Trueman 1954). Not only are coarse sediments more difficult for Softshell Clams to move through, but they are likely more difficult for crabs to dig in. Lipcius and Hines (1986) found the E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 500 Softshell Clam’s predation risk from Blue Crabs was significantly lower in sand compared to mud. Furthermore, the impact of Softshell Clam density on Blue Crab predation rates varied significantly by substrate; specifically, crab predation was density-dependent in sand, but inversely density-dependent in mud (Lipcius and Hines 1986). Softshell Clams are known to increase their burrow depth in response to chemical cues created by Green Crabs (Flynn and Smee 2010, Whitlow et al. 2003). However, this response has only been demonstrated in mud substrates and not in the whole range of sediment types in which Softshell Clams live. In this study, we examined (1) how the inducible antipredator responses of Softshell Clams are affected by sediment type (mud, sand, and gravel) and (2) how the costs and benefits (metabolic costs and predation rates) varied among sediment types. Methods Experiment no. 1: burrowing depth response Site description. Field experiments were conducted in the outer Bay of Fundy, at the Bowdoin Scientific Station on Kent Island, NB, Canada (44°35'N, 66°46'W). All organisms were collected from the intertidal zone on the western half of Kent Island. The experimental site was located on a sandy intertidal flat between Kent Island and Hay Island. All fieldwork was conducted between 1 June and 27 August 2010. Experimental set-up. Thirty plastic storage containers (46.7 x 59.4 x 36.6 cm) with small holes in the bottom were dug into the substrate such that the tops of the containers were as flush with the surrounding substrate as possible. Ten transects of 3 containers each were placed perpendicular to the shoreline (Fig. 1) in the lower intertidal zone, with an average exposure of 2–2.5 hours per tidal cycle. Thirty centimeters separated the containers within each transect. The transects were separated by 4 m. One container in each group was filled with sand, one was filled with mud, and one with gravel. Mud was collected from a sheltered salt marsh bay near the study site. Sand was collected from the site. Gravel was collected from the upper intertidal area of the study site, sifted through a 1- x 1-cm mesh to remove large rocks, and then mixed with sand in a 1 to 1 ratio by volume. Figure 1. Experimental set-up for Experiments No. 1 and 2. Black circles represent caged crabs and open circles represent empty cages. Boxes symbolize containers filled with (M) mud, (S) sand, or (G) gravel. 501 E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 Samples were taken for each sediment type used and characterized by percent weight using the Wentworth scale (Table 1; Wentworth 1922). Mud, sand, and gravel treatments were placed in random order within each transect; however, pairs of transects were filled in the same order to create matching control and treatment transects. The crab-exposure treatment was based on methods by Flynn and Smee (2010). Cylindrical cages 20 cm in diameter and 23 cm tall were made out of 2-mm x 3-mm plastic mesh. The bottom ≈8 cm of each cage was dug into the sediment, and the inside was filled with sediment up to 8 cm to anchor the cage. Cages were placed in between sediment containers and at the ends of each transect so that 2 cages surrounded each container and 4 cages were in each transect of containers. One Green Crab and a small piece of rockweed (Fucus spp.; for shelter during low tide) were placed in cages surrounding experimental treatment transects. Cages surrounding control treatments contained only a small piece of rockweed. Organism collection and care. Before being placed in the experimental plots, Softshell Clams were collected from sandy flats surrounding the experimental site. Clams ranged in size from 30–62 mm with a mean size of 46.7 mm and were randomly assigned to treatments. No significant difference in clam size distribution was found between any of the treatment groups (Kolmogorov-Smirnov Test: P > 0.05 for all, n = 20, K = 6). Prior to being placed in experimental plots, clams were measured, fastened to a tether, and uniquely marked (see below). This process took up to 24 hours during which the clams were stored in a floating lobster crate (100 × 75 × 45 cm) filled with rockweed that was moored in the lower intertidal zone. Green Crabs were collected from the rocky intertidal zone above the clam collection site. All crabs collected were male, had a carapace width of 6.0 to 8.4 cm, and had no missing claws or legs. Because it took several days to collect enough Green Crabs for each trial, crabs were kept 3–4 days in a lobster crate filled with rockweed that was tied to a mooring in the intertidal zone before being placed in cylindrical cages around the sediment plots. While they were being held temporarily, the crabs were fed Softshell Clams. Burrowing depth measurement. The end of a 60-cm monofilament string was glued to the center of each clam’s right valve (when looking at the hinge with the siphon pointing up), and a knot was tied exactly 30 cm from the clam (Griffiths Table 1. Sediment treatments from experiments no. 1 and 2 characterized by percent weight using the Wentworth scale. Very coarse to Fine to very Gravel coarse sand Medium sand fine sand Mud (>2000 μm) (500–2000 μm) (250–500 μm) (63–250 μm) (<63 μm) Mud treatment 2% 20% 17% 43% 18% Sand treatment 0% 33% 42% 25% 0% Gravel treatment 46% 25% 19% 10% 0% E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 502 and Richardson 2006). At the beginning of the experiment, clams were placed foot-down at the sediment surface, with the top of their valves even with the top of the sediment. To prevent predation, containers were covered with 1- x 1-cm plastic mesh (Flynn and Smee 2010). Meshes were cleared of algae and rockweed daily. Clam depths were measured every other day for 8 days by gently pulling the tether taut, measuring the distance between the knot and the sediment surface, and subtracting from 30 cm. Clam depths were measured again after 33 days. Analysis: experiment no. 1. A single sediment container with four clams in it was treated as one unit of replication. Mean burrowing depth was calculated for each container on every day depth measurements were taken. Using PASW Statistics 18 (IBM SPSS; Somers, NY), a repeated measures two-way ANOVA was run on the mean burrowing depths in the mud, sand, and gravel treatments over the 33 days of the experiment to determine the effect of time, crab presence, and sediment type on burrowing depth. Mauchly’s test of sphericity resulted in P < 0.05, so sphericity could not be assumed. Therefore, results were reported based on the Greenhouse-Geisser test to correct for this violation (Greenhouse and Geisser 1959). All measurements in each experiment were expressed as mean ± SE. Experiment no. 2: burrowing depth response costs Site description, experimental set-up, and organism collection and care. The same site, experimental set-up, and organisms were used in experiment no. 2 as in experiment no. 1. Experiments no. 1 and no. 2 were conducted simultaneously. Growth measurements. Valve width and length were measured using calipers prior to placing clams in the experimental plots. The right valve was measured for all length measurements. After 33 days, clams were removed and the final length of each clam, as well as the dry weights of the shell and soft tissue, were measured. Analysis: experiment no. 2. Shell thickness (log shell weight/log shell length) and percent shell weight (arcsin(dry shell weight/total dry weight)1/2) were calculated for each clam (Newell and Hidu 1982). The mean shell thickness and percent shell weight for clams in each sediment container were calculated and compared with a two-way ANOVA using PASW Statistics 18 (IBM SPSS; Somers, NY) to determine the effect of sediment type and crab presence on growth. Statistical analysis on transformed and untransformed percent shell weights were in agreement; so analysis of untransformed (dry shell weight/total dry weight) values are reported for simplicity (Newell and Hidu 1982). Experiment no. 3: predation risk Site description. Lab experiments were conducted in tanks at Bowdoin College’s marine laboratory in Harpswell, ME (43°47'N, 69°57'W) from 20 October to 11 December 2010. The laboratory houses large tanks with flow-through seawater pumped from Harpswell Sound. Experimental set-up. Twelve rectangular plastic storage containers (34 x 45 x 30 cm) were filled with 7, 9, 11, or 13 cm of mud, sand, or gravel. There 503 E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 was one container holding each depth × sediment combination. Mud was collected from the intertidal zone near the marine lab. Playground sand and pea gravel were purchased from Lowe’s Hardware (Brunswick, ME). The sediment types were characterized as in experiment no. 1 (Table 2). Sediment containers were placed in large flow-through seawater tanks (180 x 94 x 50 cm), which were filled with approximately 40 cm of water so that all the sediment containers were completely submerged in water. Two clams were placed at the bottom of each sediment container with their siphons upwards. Clams in preliminary trials successfully extended their siphons to the sediment surface and remained buried at the bottom of the containers. Furthermore, upon removal of clams at the end of each trial, it was confirmed that clams were still located at the bottom of the containers. Organism collection and care. Green Crabs and Softshell Clams were collected by hand from the intertidal zone in Harpswell Sound. Clam lengths ranged from 40 to 49.9 mm, and crabs had carapace widths of 42 to 55 mm. Male and female crabs with both claws and all of their legs intact were used. Organisms were held for up to 6 weeks in flow-through seawater tanks at the Bowdoin marine lab prior to use. During the holding period prior to their experimental trials, crabs were provided clams to eat. However, food was withheld from the crabs for six days prior to the beginning of the experiment. Crab predation. One crab was randomly placed in each of the containers, which were covered with a rigid plastic grate to prevent crabs from escaping. The number of clams eaten was recorded after 6 days. At the end of a six-day trial, new clams and crabs were placed in the containers, so that each organism was used in only one trial, and the experiment was repeated for a total of 6 replicates. Tanks received constant flow-through seawater for the first 3 cycles. However, crab activity decreased with decreasing water temperature as autumn progressed into winter. So the tank configurations were changed to re-circulate water, which warmed the water to ≈12–13 °C for the final 3 cycles. For these final three cycles, the water was changed before the start of each new cycle. Predation trials were tested to ensure results did not vary significantly by circulation method, recirculating versus flow-through water, in any of the substrates (Kruskal-Wallis: mud P = 0.598, sand P = 0.514, gravel P = 0.265; n = 12, K = 2). Analysis: experiment no. 3. Using Prism 5 Version 5.0a (GraphPad Software, Inc; La Jolla, CA), Kruskal-Wallis tests were performed comparing the mean percentage of clams eaten at each depth and in each sediment type. Table 2. Sediment treatments from experiment no. 3 characterized by percent weight using the Wentworth scale. Very coarse to Fine to very Gravel coarse sand Medium sand fine sand Mud (>2000 μm) (500–2000 μm) (250–500 μm) (63–250 μm) (<63 μm) Mud treatment 8% 22% 13% 40% 17% Sand treatment 0% 17% 41% 41% 1% Gravel treatment 48% 10% 21% 20% 1% E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 504 Results Experiment no. 1: burrowing response Time, crab presence, and sediment type had significant effects on the depths to which Softshell Clams burrowed (repeated measures two-way ANOVA: time P < 0.001, crab presence P = 0.014, sediment type P < 0.001, interaction P = 0.817; n = 5, K = 6; Fig. 2). Clams exposed to crab cues tended to burrow to deeper depths (depth on day 33 = 13.1 ± 1.8 cm) in all sediment types when compared to clams that were not exposed to crab cues (10.3 ± 1.3 cm). Furthermore, the difference between burrowing depths of crab-exposed clams and control clams increased over time. Clams in the mud and gravel treatments reached a maximum difference in depth between crab-exposed clams and control clams at the end of the experiment on day 33 (depth on day 33: mud with crabs = 14.4 cm ± 2.3 cm, mud with no crabs = 10.6 cm ± 0.9 cm, sand with crabs = 12.9 cm ± 1.1 cm, sand with no crabs = 11.0 cm ± 1.6 cm, gravel with crabs = 12.0 cm ± 1.0 cm, gravel with no crabs = 9.3 cm ± 0.9 cm). However, clams in sand reached a maximum difference in depth between crab-exposed clams and control clams on day 8, after which the difference decreased slightly (depth ± SE on day 8: mud with crabs = 11.3 cm ± 1.7 cm, mud with no crabs = 8.3 cm ± 1.1 cm, sand with crabs = 13.6 cm ± 0.9 cm, sand with no crabs = 11.3 cm ± 1.0 cm, gravel with crabs = 11.5 cm ± 1.7 cm, gravel with no crabs = 9.6 cm ± 1.1 cm). Still, the difference between crab-exposed and control clams in sand treatments on day 8 was smaller than those in mud and larger than those in gravel, as predicted. Throughout the experiment, crab-exposed clams in the mud treatments had the deepest mean burrow depths as well as the greatest difference in burrow depth between clams exposed and not exposed to crab cues. Experiment no. 2: cost of burrowing response Shell thickness increased with increasing grain size of the sediment (twoway ANOVA: crab presence P = 0.244, sediment type P = 0.018, interaction P = 0.111; n = 5, K = 6; Fig. 3A). Shells of clams in mud were significantly thinner than were those of clams in gravel treatments (mud = 0.298 ± 0.044 mm, sand = Figure 2. Mean burrow depths of Softshell Clams in (A) mud, (B) sand, and (C) gravel after 33 days. Shaded bars represent crab-exposure treatments, and open bars represent controls. Repeated measures two-way ANOVA: time P < 0.001, crab presence P = 0.014, sediment type P < 0.001, interaction P = 0.817; n = 5, K = 6. Error bars represent ±1 standard error of the mean. 505 E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 0.311 ± 0.027 mm, gravel = 0.344 ± 0.036 mm; Tukey post hoc: mud vs. gravel P = 0.017, all other comparisons P > 0.05; n = 10, K = 3). Furthermore, clams exposed to crabs in mud treatments had significantly thicker shells (0.33 ± 0.03 mm) than those of control clams (0.27 ± 0.04 mm) in mud treatments (t-test: mud P = 0.0414, sand and gravel P > 0.05; n = 5, K = 2). Mean percent shell weight varied significantly by sediment (two-way ANOVA: crab presence P = 0.699, sediment type P = 0.021, interaction P = 0.574; n = 5; Fig. 3B). Clams in gravel treatments had proportionately heavier shells than did clams in sand (mud = 0.867 ± 0.001, sand = 0.863 ± 0.002, gravel = 0.874 ± 0.009; Tukey post hoc: sand vs. gravel P = 0.017, all other comparisons P > 0.05; n = 5, K = 6). Although not statistically significant, clams living in coarser sediments trended towards having the smallest increase in length. Furthermore, mean change in length of crab-exposed clams was smaller than that of control clams within each sediment type. Experiment no. 3: predation risk The mean percentage of clams eaten varied significantly by sediment type, but not by burrowing depth (Kruskal-Wallis: depth P = 0.592, n = 18, K= 4; sediment type P = 0.002, n = 24, K = 3; Fig. 4). Clams in mud were most vulnerable (percent preyed on: mud = 0.40 ± 0.42, sand = 0.08 ± 0.24, gravel = 0.10 ± 0.26; Dunn’s multiple comparison post hoc: mud vs. sand and gravel both P < 0.01, all others P > 0.05; n = 24, K = 3). Very few clams in sand and gravel were preyed upon. Furthermore, no clams in the deepest sand and gravel treatments were preyed on, while clams in all depth treatments of the mud sediment were preyed on. Figure 3. (A) Mean shell thickness and (B) percent shell weight of Softshell Clams in 3 sediment types after 33 days. Shaded bars represent crab-exposure treatments and open bars represent controls. Two-way ANOVA: n = 5, K = 6. (A) crab presence P = 0.244, sediment type P = 0.018, interaction P = 0.110. (B) crab presence P = 0.699, sediment type P = 0.021, interaction P = 0.574. Error bars represent ±1 standard error of the mean. E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 506 Discussion When exposed to Green Crabs, Softshell Clams increased their burrowing depths in all sediment types. However, Softshell Clams burrowed more deeply in mud than in the coarser sediment treatments. Clams in gravel experienced relatively low risk of predation by crabs compared to those living in finer sediments. But clams in gravel appeared to incur higher locomotive and maintenance costs due to the more abrasive coarse sediment. Shell thickness was positively correlated with sediment grain size, whereas growth rate was negatively correlated with grain size. Sediment grain size significantly affected the magnitude of Softshell Clams’ burrowing response; the strength of the response increased with decreasing sediment grain size (Fig. 2). Previous studies suggested burrowing in coarse sediments is more costly than in fine ones. Trueman (1954) found that sand was more resistant to shell opening (which is required for burrowing) than was mud and that burrowing becomes increasingly difficult as Softshell Clams get deeper in the sediment. This finding suggests that in coarse sediments, where it is already difficult to gape at shallow depths, burrowing deep into the sediment may be quite costly. Growth in length of Softshell Clams tended to be negatively correlated with sediment grain size, suggesting that it is more costly to inhabit coarse sediments. Brousseau and Baglivo (1988), Newell and Hidu (1982), and Swan (1952) also found Softshell Clams living in coarse sediments had lower length growth rates, which they attributed to increased cost of locomotion and higher maintenance costs due to increased shell-chipping and greater difficulty in maintaining a normal gape. In our study, clams in coarser sediments had thicker shells, and their shells comprised a greater percentage of their total weight than the shells of clams in fine sediments (Fig. 3). However, previous studies (Freeman 2007) found that unlike species such as Blue Mussels, which increase shell thickness as an antipredator Figure 4. Rate of predation by Green Crabs on Softshell Clams burrowed to various depths in (A) mud, (B) sand, and (C) gravel. Kruskal-Wa l l i s : depth P = 0.592, n = 18, K= 4; sediment type P = 0.002, n = 24, K = 3. Error bars represent ±1 standard error of the mean. 507 E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 response, increased shell thickness in Softshell Clams would be ineffective in reducing predation risk due to its gape. Experiments with Blue Crabs showed that no size class of Softshell Clams was immune from predation (Blundon and Kennedy 1982a). Thus, due to its gape and brittle shell, increased shell thickness does not provide any protection for Softshell Clams. Increased shell thickness is likely due to needs for greater strength against the destructive forces caused by coarse sediments. This hypothesis is supported by the larger investment in shell versus soft tissue we found for clams living in gravel, regardless of crab exposure, compared to clams in mud or sand (Fig. 3B). The results therefore suggest coarse sediment is more costly to live in due to increased maintenance costs and shell thickness requirements. Increased shell thickness was also significantly greater in clams exposed to crabs compared to clams not exposed to crabs within the mud treatment (Fig. 3A). This finding may be the result of shell regrowth following shell erosion caused by increased burrowing activity, with the regrown shell being thicker than the original shell. Together, these results suggest increased shell thickness was caused by shell regrowth stimulated by shell loss caused by inhabiting coarse sediment and by increased burrowing activity in response to the presence of predators. Predation by Green Crabs was affected by sediment type (Fig. 4), in a manner similar to that documented for Blue Crabs in mud and sand (Arnold 1984, Lipcius and Hine 1986). This pattern of decreased predation in coarser sediment found in these earlier studies extends to coarser gravel sediments for Green Crabs. Just as coarser sediments are more difficult for clams to move through, they are likely more difficult for crabs to dig in. Clams in mud were significantly more vulnerable to predation than were clams in sand or gravel. Clams in sand and gravel showed a general trend of decreased predation risk with increasing burrow depth (Fig. 4). However, no such trend was observed in clams in mud treatments; rather, clams were frequently preyed upon regardless of burrowing depth, demonstrating the high risk of crab predation for clams in mud. Our results may have been affected by the laboratory conditions in which the depth to which each clam could burrow was limited and the crabs were confined to a small space immediately above the clams. Our experimental results support the hypothesis that environmental factors such as sediment grain size, impact growth, energy allocation, and risk of predation for Softshell Clams. While coarse sediments provide the greatest survivorship due to low predation, they are the most costly to live in. Furthermore, cost of predator avoidance varied by sediment. Responding to crab presence by burrowing deeper in the sediment was the most costly in coarse sediments. These results suggest the plastic antipredator responses of Softshell Clams depend not only on temporal variation in predator density, but also on environmental factors such as sediment grain size. Plasticity of antipredator responses related to sediment type would result in spatial variation in the magnitude of the response to predators. These results also suggest that a temporal mismatch exists in the costbenefit tradeoff of Softshell Clam’s antipredator responses, with the probability of present mortality being weighed against future reproductive success. E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 508 Acknowledgments We wish to thank the students and staff of the Bowdoin Scientific Station for help with logistics. Andy Bell, Allison Chan, Grace Hyndman, Laura Newcomb, and Mark Murray helped with laboratory experiments. Cathryn Field assisted with sediment characterization. Amy Johnson and John Lichter provided feedback on the analysis and writing. E.Thomson was supported by a Bowdoin Scientific Station Summer Fellowship. Additional funding was provided by the Bowdoin College Department of Biology. This research was conducted under scientific collecting permits no. 322694 (issued by the Canadian Department of Fisheries and Oceans) and 2010-41-01 (issued by the Maine Department of Marine Resources). This is contribution number 245 of the Bowdoin Scientific Station. Literature Cited Appleton, R.D., and A.R. Palmer. 1988. Water-borne stimuli released by predatory crabs and damaged prey induce more predator-resistant shells in marine gastropod. Evolution 85:4387–4391. Arnold, W.S. 1984. The effects of prey size, predator size, and sediment composition on the rate of predation of the Blue Crab, Callinectes sapidus Rathbun, on the Hard Clam, Mercenaria mercenaria (Linné). Journal of Experimental Marine Biology and Ecology 80(3):207–219. Auffrey, L.M., S.M. Robinson, and M.A. Barbeau. 2004. Effect of green macroalgal mats on burial depth of Soft-shelled Clams, Mya arenaria. Marine Ecology Progress Series 278:193–203. Beal, B.F., M.R. Parker, and K.W. Vencile. 2001. Seasonal effects of intraspecific density and predator exclusion along a shore-level gradient on survival and growth of juveniles of the Soft-shell Clam, Mya arenaria L., in Maine, USA. Journal of Experimental Marine Biology and Ecology 264:133–169. Blundon J.A., and V.S. Kennedy. 1982a. Mechanical and behavioral aspects of Blue Crab, Callinectes sapidus (Rathbun), predation on Chesapeake Bay bivalves. Journal of Experimental Marine Biology and Ecology 65(1):47–65. Blundon J.A., and V.S. Kennedy. 1982b. Refuges for infaunal bivalves from Blue Crab, Callinectes sapidus (Rathbun), predation in Chesapeake Bay. Journal of Experimental Marine Biology and Ecology 65:67–81. Boulding, E.G. 1984. Crab-resistant features of shells of burrowing bivalves: Decreasing vulnerability by increasing handling time. Journal of Experimental Marine Biology and Ecology 76(3):201–223. Brousseau, D.J. 1979. Analysis of growth rate in Mya arenaria using the Von Bertalanffy equation. Marine Biology 51(3):221–227. Brousseau, D.J., and J.A. Baglivo. 1988. Life tables for two field populations of Softshell Clam, Mya arenaria (Mollusca: Pelecypoda), from Long Island Sound. Fishery Bulletin 86(3):567–579. Carriker, M.R. 1959. The role of physical and biological factors in the cultures of Crassostrea and Mercenaria in a salt-water pond. Ecological Monographs 29(3):219–266. Commito, J.A. 1982. Effects of Lunatia heros predation on the population dynamics of Mya arenaria and Macoma balthica in Maine, USA. Marine Biology 69:187–193. 509 E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 Dill, L.M., and J.F. Gillett. 1991. The economic logic of barnacle Balanus glanduia (Darwin) hiding behavior. Fishery Bulletin 153:115–127. Floyd, T., and J. Williams. 2004. 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. Journal of Shellfish Research 23(2):457–462. Flynn, A.M., and D.L. Smee. 2010. Behavioral plasticity of the Soft-shell Clam, Mya arenaria (L.), in the presence of predators increases survival in the field. Fishery Bulletin 383:32–38. Freeman, A.S. 2007. Specificity of induced defenses in Mytilus edulis and asymmetrical predator deterrence. Marine Ecology Progress Series 334:145–153. Glude, J.B. 1955. The effects of temperature and predators on the abundance of the Softshell Clam, Mya arenaria, in New England. Transactions of the American Fisheries Society 84:13–26. Gosner, K. 1987. Peterson Field Guides: A Field Guide to the Atlantic Seashore. Houghton Mifflin Company, Boston, MA. 352 pp. Greenhouse, S.W., and S. Geisser. 1959. On methods in the analysis of profile data. Psychometrika 24(2):95–112. Griffiths, C.L., and C.A. Richardson. 2006. Chemically induced predator avoidance in the burrowing bivalve Macoma balthica. Journal of Experimental Marine Biology and Ecology 331:91–98. Harvell, C.D. 1990. The ecology and evolution of inducible defenses. The Quarterly Review of Biology 65(3):323–340. Holland, A.F., N.K. Mountford, M.H. Hiegel, K.R. Kaumeyer, and J.A. Mihursky. 1980. Influence of predation on infaunal abundance in Upper Chesapeake Bay, USA. Marine Biology 57(3):221–235. Kamermans, P. 1994. Similarity in food source and timing of feeding in deposit- and suspension-feeding bivalves. Marine Ecology Progress Series 104:63–75. Lipcius, N.R., and A.H. Hines. 1986. Variable functional responses of a marine predator in dissimilar homogeneous microhabitats. The Ecological Society of America 67(5):1361–1371. Matthiessen, G.C. 1960. Observation on the ecology of the Soft Clam, Mya arenaria, in a salt pond. Limnology Oceanography 5(3):291–300. Newell, C.R., and H. Hidu. 1982. The effects of sediment type on growth rate and shell allometry in the Soft-shelled Clam, Mya arenaria. Journal of Experimental Marine Biology and Ecology 65:285–295. Ropes, J.W. 1968. The feeding habits of the Green Crab, Carcinus maena (L.). Fishery Bulletin 67(2):183–203. Rowell, T.W., and P. Woo. 1990. Predation by the nemertean worm, Cerebratulus lacteus Verrill, on the Soft-shell Clam, Mya arenaria Linneaus, 1758, and its apparent role in the destruction of a clam flat. Journal of Shellfish Research 9(2):291–297. Sih, A. 1987. Prey refuges and predator-prey stability. Theoretical Population Biology 31:1–12. Smee, D.L., and M.J. Weissburg. 2006. Hard Clams (Mercenaria mercenaria) evaluate predation risk using chemical signals from predators and injured conspecifics. Journal of Chemical Ecology 32(3):605–619. Smith, T.E., R.C. Ydenberg, and R.W. Elner. 1999. Foraging behavior of an excavating predator, the Red Rock Crab (Cancer productus Randall,) on Soft-shell Clam (Mya arenaria L.). Journal of Experimental Marine Biology and Ecology 238:185–197. E. Thomson and D.P. Gannon 2013 Northeastern Naturalist Vol. 20, No. 3 510 Swan, E.F. 1952. The growth of the clam Mya arenaria as affected by the substratum. Ecology 33(4):530–534. Taylor, D.L., and D.B. Eggleston. 2000. Effects of hypoxia on an estuarine predator-prey interaction: Foraging behavior and mutual interference in the Blue Crab, Callinectes sapidus, and the infaunal clam prey Mya arenaria. Marine Ecology Progress Series 196:221–237. Trueman, E.R. 1954. Observations on the mechanism of the opening of a burrowing lamellibranch, Mya arenaria. Journal of Experimental Biology 31:291–305. Welch, W.R. 1968. Changes in abundance of the Green Crab, Carcinus maenas (L.), in relation to recent temperature changes. Fishery Bulletin 67(2):337–345. Wentworth, C.S. 1922. A scale of grade and class terms for clastic sediments. The Journal of Geology 30(5):377–392. Whitlow, W.L., N.A. Rice, and C. Sweeney. 2003. Native species vulnerability to introduced predators: Testing an inducible defense and a refuge from predation. Biological Invasions 5:23–31. Zaklan, S.D., and R. Ydenberg. 1997. The body size-burial relationship in the infaunal clam Mya arenaria. Journal or Experimental Marine Biology and Ecology 215:1–17.