Determining the ecological role of Euspira lewisii: Part I: feeding ecology.ABSTRACT The predatory naticid snail Euspira lewisii, Lewis' moon snail, native to the west coast of North America, is stated to be an economic threat to the shellfish aquaculture industry in British Columbia (BC). This species is being manually removed from the intertidal ecosystem, yet little is known about the ecology of this species. Enclosures and beach shell assemblages were used to determine the prey preference, feeding rates, and community impacts orE. lewisii. In the enclosures, Protothaca staminea, the native little neck clam, was found to be the preferred prey, whereas the commercially valuable Manila clam, Venerupis philippinarum, was avoided when E. lewisii was offered other clam prey choices. Drilled shells collected from the intertidal revealed similar feeding preferences. The feeding rate averaged for a variety of clam species was found to be 0.09 clams/day or 1 clam every 14 days. The feeding rate was dependent on prey species, and was highest for the preferred species and significantly lower on avoided species. The overall impact of E. lewisii to the bivalve community was found to be low. Based on these results, E. lewisii consumed only approximately 3% of the clam population over 1 y, assuming maximal feeding rates and typical population densities found on the west coast of BC. E. lewisii has minimal impacts to the Manila clam industry in BC and control measures are not necessary for this species. Baseline ecological field studies are important for gaining an understanding of poorly understood species, especially those considered threats to industry.
KEY WORDS: British Columbia, community impacts, gastropod, Euspira lewisii, feeding ecology, shellfish aquaculture
Predation is one of the most important factors affecting community structure in intertidal communities. It can affect the distribution pattern, size and age composition, and abundance of prey species (Peitso et al. 1994, Beal 2006). Recent studies have stressed the importance of a full understanding of predation such that we can evaluate the ecological impact a predator has on a community (Savini & Occhipinti-Ambrogi 2006). The key to understanding the role a predator plays in a community includes knowing its prey preferences and feeding rates with respect to the densities of both the predator and the prey (Moran 1985, Thiel et al. 2001). From an applied aspect, without a full understanding of predation, it is difficult to manage intertidal communities or know whether antipredator practices, such as predator removal, are effective (Miron et al. 2005).
Many intertidal predators demonstrate prey preferences and select prey that maximize net energy intake by having a caloric value higher than the energy costs of capturing, handling, and consuming that prey item (Savini & Occhipinti-Ambrogi 2006). The effects of selective predation on community structure vary with relative abundance of prey species (Moran 1985), the escape abilities of the prey species and the nature of the habitat. Selective feeding on nondominant species can have adverse effects on unstable communities, such as decreasing species diversity by removing rare species (Wiltse 1980a).
Feeding rates of predators depend on a number of biotic and abiotic factors. Biotic factors include prey biomass, density, life cycle, species, quality, and predator and prey size (Wiltse 1980b, Commito 1982, Moran 1985, Thiel et al. 2001, DeGraaf & Tyrrell 2004). Time spent on other activities such as mating, competition, or predator avoidance also influences feeding rates (Thiel et al. 2001, DeGraaf & Tyrrell 2004). Abiotic factors such as temperature, season, wave action, and duration of submersion (Moran 1985, Weissberger 1999) also affect feeding rates. Greater feeding rates can lead to greater impacts on the prey community, such as reduced abundance of the prey species (DeGraaf & Tyrrell 2004, Savini & Occhipinti-Ambrogi 2006).
Euspira lewisiiis a gastropod in the family Naticidae, a group commonly known as moon snails, which are infaunal, predatory snails that feed on bivalves. This species can be found crawling at the surface of intertidal sediments or within the sediments to depths up to 20cm (Bernard 1967). Several species of moon snails have shown both size and species preferences while feeding (Bernard 1967, Wiltse1980a; Commito 1982, Rodrigues et al. 1987, Peitso et al. 1994, Dietl & Alexander 1997). Through drilling activities, very clear evidence of the predation of these species is left in intertidal habitats, most commonly on the Pacific littleneck clam (Protothaca staminea), the butter clam (Saxidomus gigantea), and the Pacific horse clam (Tresus nuttallii). For this reason, they are considered pest species, especially to Manila clam shellfish aquaculture (Bernard 1967, Peitso et al. 1994, British Columbia Shellfish Growers Association 2002, Beal 2006). However, little is known about the predation pressure of the moon snail on bivalve populations. A review of the literature suggests that bivalve mortality attributed to moon snails may, in fact, be overestimated (Wiltse 1980b, Peitso et al. 1994, Beal et al. 2001, Miron et al. 2005).
Manila clams, Venerupis philippinarum, is the primary cultured clam species in British Columbia (BC) and makes up 46% of BC shellfish aquaculture by value (BC Ministry of the Environment, Oceans and Marine Fisheries Branch 2008). The Manila clam industry brings in a total revenue of $192 million per year (Department of Fisheries and Oceans 2006).
On the west coast of BC, the native moon snail E. lewisii is being destroyed and eliminated from Manila clam shellfish farms, based on the assumption that they are effective predators. Hence, the objectives of this study are to assess the impacts of predation by E. lewisii on bivalve communities with special emphasis on the commercially valuable Manila clam, V. philippinarum. We use both field experiments and the collection of drilled bivalve shells to determine moon snail prey preference, feeding rates, and impacts on the prey community. Based on previous studies, we predict that E. lewisii will show distinct prey preferences, their feeding rate will be lower than previously thought, and thus their overall impact on bivalve communities will be relatively low.
MATERIALS AND METHODS
Field research was conducted intertidally in southern BC at Fillongley Provincial Park, on Denman Island (49[degrees]31'59"N, 124[degrees]49'0"W) and Shingle Spit, on Hornby Island (49[degrees]31'0"N, 124[degrees]37'59"W). The sites were selected because they were adjacent to Manila clam aquaculture leases and so were similar in terms of sediment characteristics and species composition (Whiteley & Bendell-Young 2007). Both sites are home to known populations of E. lewisii where they were found buried at depths from directly beneath the surface to approximately 10 cm, some actively drilling bivalve prey (pers. obs.) V. philippinarum, the commercially valuable Manila clam, and P. staminea, the native Pacific littleneck clam, dominate the bivalve community at these sites. Nuttallia obscurata, the varnish clam, a recent introduction to southern BC, as well as several other clam species are also found at these sites.
We used enclosure experiments (12 cages) to determine the prey preferences of E. lewisii. Each cage was made of a PVC pipe frame measuring 1 X 1 X 0.3 m and was enclosed in an area of 1 [m.sup.2]. All sides of the frame were covered with plastic mesh with an aperture of 1 [cm.sup.2]. The cages were dug into the sediment to a depth of 0.2 m, leaving 0.1 m exposed at the surface. Sediment was returned to the cage after all bivalves and drilled shells were removed. A 4 x 3 grid was created, using 12 cages, oriented parallel to the water line (Fig. 1). The cages in the grid were spaced approximately 2 m apart and were located between 1.7 m and 1.4 m above chart datum to incorporate the distributions of all the species and were placed at tidal height where shellfish tenures are located, which is between 1.5 m and 2.5 m above chart datum (Department of Fisheries and Oceans 2006).
Three clam species collected from Fillongley were used in the experiments to analyze the prey preferences of E. lewisii: P. staminea, V. philippinarum, and N. obscurata. Twenty clams of each species were buried in each cage, with 5 individuals of each species in each corner. The clams were selected to be more than 38 mm, the legal size of P. staminea and V. philippinarum, and within the preferred prey size of E. lewisii based on the findings of Peitso (1980). N. obscurata was also selected in this size range to eliminate the possibility of prey preference resulting from size. This led to 60 clams in each cage (60/[m.sup.2]) and 720 clams in all 12 cages. This density was selected because, in the area, the clam density was found to be 70 clams/[m.sup.2] and it is easiest to divide 60 into the three species of interest. This density is less than on aquaculture leases, but Peitso (1980) found that clam density did not have a significant effect of E. lewisii feeding rates. Two cages, selected at random and containing clams and no moon snails, served as controls that tested for clam transplant survivorship. In the 10 remaining cages, a single moon snail, collected from the site, was measured and buried into the center of the cage. Peitso et al. (1994) found that E. lewisii between 72.4 mm and 95 mm in length fed preferentially on clams between 35 mm and 50 mm shell length. Therefore, the snails we selected were between these sizes. The E. lewisii density was representative of the study areas. All cages were sealed and left.
[FIGURE 1 OMITTED]
The cages were checked every other tide cycle, approximately once every 3 wk throughout the course of 4 mo, and all drilled and dead clams were removed and replaced with live individuals of the appropriate species and size. Only completely drilled shells were used in the prey preference analyses. Very few (n = 6) clams died without drill marks, and those that were not drilled died of exposure at the surface so were not included in calculations. The prey preference experiments were run from May through September, summer 2005 and 2006.
Manly's [alpha] was used as an index of preference for constant prey populations (see Krebs 1999):
[[alpha].sub.i] = [r.sub.i]/[n.sub.i](1/[summation] ([r.sub.j]/[n.sub.j]))
where [[alpha].sub.i] is Manly's [alpha] (preference index) for prey type i; [r.sub.i[, [r.sub.j] is the proportion of prey type i or j in the diet (i and j = 1,2,3, ... m); [n.sub.i], [n.sub.j] is the proportion of prey type i or j in the environment; and m is the number of prey types possible.
Similar preference experiments have used this index (Dudas et al. 2005), and it is well established in the feeding preference literature (Manly et al. 1972, Manly 1974, Krebs 1999). The interpretation of the [alpha] values for this index are
[[alpha].sub.i] = 1/m = no preference for species i
[[alpha].sub.i] > 1/m = preference for species i
[[alpha].sub.i] < 1/m = avoidance of species i
where m = number of prey species.
For these experiments, 3 species were used; therefore, an [alpha] value of 0.33 indicates no preference, more than 0.33 is an indication of preference, and less than 0.33 is an indication of avoidance. These ct values are considered significant if the 95% confidence intervals does not overlap the 0.33 prey types. A log-likelihood chi-square analysis was performed to substantiate these results.
Feeding rates were determined in tandem with the prey preference data. Feeding rates were calculated as the number of clams consumed divided by the number of days the moon snail was contained within the cage, and as the number of days the moon snail was contained within the cage divided by the number of clams consumed. An average of each of these measures was then taken.
By-species feeding rates were also determined. In summer 2006, a trial was carried out in which the 12 cages were randomly selected to contain one of each of the 3 species. Fifteen individuals of each species were buried in each of the 4 corners of the cage (i.e., 4 cages per species, 240 clams per species, for a total of 720 clams). Snails were added as described above and the cages were sealed for approximately 3 wk. After the 3 wk, all the cages were checked, and any drilled shells were removed and tallied. This design was representative of a monoculture situation similar to those observed on the nearby clam tenures, thus observations could be made on how E. lewisii might feed under the conditions observed on the leases.
The Kruskal-Wallis test was applied to determine significant differences among species feeding rates on these 3 clam species because the data were not normally distributed.
Density and Drill Collection
Density surveys of all bivalve species and moon snails were conducted at both the Fillongley Provincial Park site and at Shingle Spit. To account for tidal influences, a 60-m-wide strip parallel to the shoreline representative of the intertidal communities was stratified into tide heights by dividing it into high, mid, and low zones (Table 1). The zones were different in size as a result of the geomorphology of the beach at each site.
Within each stratum, four 60-m-long transects parallel to the shore were randomly selected at Fillongley and three 60-m-long transects were randomly selected at Shingle Spit. The beach at Shingle Spit was steeper than that at Fillongley. Along each transect, 6 quadrat locations were selected at random. This led to 24 quadrats sampled in each stratum at Fillongley for a total of 72 quadrats, whereas 18 quadrats were sampled in each stratum at Shingle Spit for a total of 54 quadrats sampled. Random numbers were selected using a random number table. At each coordinate, a 0.5 X 0.5-m quadrat was dug (0.25 [m.sup.2]) down to a depth of 0.2 m. All sediment dug from the quadrat was sifted through a 6-mm mesh, and all infaunal bivalves were identified and counted to determine community composition and densities. Whiteley (2005) found that only 10% of species and 25% of species count data were lost using 6-mm versus 1-mm sieve mesh. The larger aperture mesh also allowed for increased sampling, because field researchers were not limited by the lengthy sieving time using 1-mm mesh. This sampling was carried out from May until September for 2 summers, as tides allowed.
During the sifting process, any shells containing the distinct countersunk E. lewisii drill marks were removed and the clam species was identified (Peitso et al. 1994). All live organisms and drilled shells were replaced after sampling.
E. lewisii densities were determined using a mark-recapture technique carried out in May 2006. In an area of 660 [m.sup.2] at Shingle Spit and 2,010 [m.sup.2] at Fillongley that incorporated the high, mid, and low zones described previously, 50 individual snails were marked by scratching a number into their shell, then the snails were reburied in the sediment. After 3 wk, we returned and dug up 30 snails and determined the number of marked snails. This technique makes several assumptions: that the population is closed (no births, deaths, immigration, or emigration), the mark remains in place, all animals are equally likely to be captured in each sample and over time, and also that being trapped once does not make an individual more or less likely to be captured again. Assuming that the marked snails return homogeneously to the general population, the total E. lewisii population was calculated based on the work by Bernard (1967) as follows:
T = M/(R/C)
where T is the total population in the area, M is the number of marked animals in the 1st sample, R is the number of marked animals in the 2nd sample, and C is the total caught in the 2nd sample.
Prey Preferences from Beach Shell Assemblages
The density measurements and drills collected were used to determine whether E. lewisii prey preferences were also evident under natural conditions. Proportions of the clams were calculated based on a stratified multistage design (Krebs 2001, Schwarz 2005). The proportions of shells and species in the community were also used to calculate electivity coefficients (E) based on the work by Ivlev (1961):
E = (r - p)/(r + p)
where r is the proportion of a food item in the diet and p is the proportion of the food item in the environment. Preference is indicated by a positive value of E, avoidance is indicated by a negative value, and no preference is indicated by a value of 0.
Ivlev's electivity coefficient Ivlev (1961) was selected because of the variable nature of the bivalve communities in the intertidal. Manly's a is appropriate for constant prey populations or in experimental situations when the prey is being replaced, maintaining a constant supply of food (Krebs 1999). It is not recommended that [alpha] values be calculated based on populations with different numbers of prey types (Krebs 1999).
It is also important to consider the assumptions of using a beach shell assemblage to describe prey preference. It is assumed that all clam species stand up equally to abiotic factors such as storms, currents, and dissolution. It must also be assumed that the observed prey preference is not a result of a size refuge in the prey species. The final assumption is that the shells originated in the same location as they were collected.
Density measurements and average feeding rates were used to represent the effects of E. lewisii predation on these intertidal communities. The average feeding rate was used to calculate the number of clams consumed in a month, in 6 mo, and in 1 y based on
Number of clams consumed = (feeding rate) X (days) X (number of snails)
For these calculations we used the total number of snails in our study areas from mark--recapture estimates. Based on information from the Department of Fisheries and Oceans (2006), V. philippinarum are harvested at an average weight of 18 g and sell for $5.00/kg. This leads to the price of $0.09/clam. These values allow us to calculate the financial cost of E. lewisii predation to the Manila clam industry.
When offered equal numbers of P. staminea, V.philippinarum, and N. obscurata, E. lewisii showed significant preference for P. staminea ([alpha] = 0.57, chi-square = 51.25, P < 0.0001; Fig. 2). N. obscurata was preferred although it was not statistically significant (Fig. 2). V. philippinarum was not a significantly selected prey item ([alpha] = 0.07, chi-square = 51.25, P < 0.0001; Fig. 2).
[FIGURE 2 OMITTED]
The average summer feeding rate of E. lewisii consuming a variety of prey species was found to be 0.09 [+ or -] 0.02 clams/day ([+ or -] 95% confidence interval), with 1 clam consumed every 14 days. This is the feeding rate determined for E. lewisii.
When the feeding rates were analyzed for each of the 3 species individually, the consumption rate on P. staminea was greater than that on N. obscurata, which was greater than the rate on V. philippinarum (Fig. 3).
The Kruskal-Wallis test showed that the feeding rates in clams per day were significantly different (H = 6.17, P < 0.05; Fig. 3). The feeding rate on V. philippinarum was significantly different from that of P. staminea. N. obscurata was not significantly different from either species (Wilcoxon, P < 0.05).
Bivalve and E. lewisii Density and Abundance
Based on the transect density surveys at both sites, the total density per square meter of bivalves decreased as the tide level decreased (Fig. 4). The density of V. philippinarum was greatest in the high tide zone and decreased through the other strata to the waterline. However, at both sites the density followed the same pattern of being highest in the mid-intertidal, followed by the high zone, and was the least dense in the low zone. N. obscurata was found in very low densities in the study areas and was only found in the high zone and to a lesser extent in the mid tide zone (Fig. 4). Macoma spp. was found in much higher densities at Shingle Spit, and at both sites it was at its highest densities in the mid and low strata (Fig. 4). The other species we found at both sites were Mya arenaria, Saxidomus gigantea, Parvaleucina tenuisculpta, and Rhamphidonta retifera. Tellina carpenteri, Clinocardium nuttallii, and Lyonsia californica were exclusively found at Fillongley whereas T. nuttallii and Cryptornya californica were only found at Shingle Spit. Macoma spp. was predominantly Macoma nasuta, hut at smaller sizes it was difficult to distinguish it from Macoma obliqua, so both of these species were represented in these communities.
V. philippinarum was the most abundant species at both sites (Table 2). P. staminea was the second most abundant species at Fillongley whereas Macoma spp. was the second most abundant species at Shingle Spit (Table 2). From the quadrat surveys, E. lewisii was found only in the lowest stratum at Fillongley at an abundance of 2,000 [+ or -] 3,000 individuals in a 60-m strip of the intertidal, although data variability for these measurements was very high (Table 3). E. lewisii was collected in both the mid and the low strata at Shingle Spit, and estimates suggested abundances of 200 [+ or -] 300 snails and 300 [+ or -] 500 snails in the mid and low stratums, respectively (Table 2). E. lewisii densities were more accurately estimated using the mark-recapture techniques and were found at all 3 tide heights. After recapturing 6 marked snails out of 50 at Fillongley and 10 out of 30 at Shingle Spit, the density of E. lewisii was 0.2 snails/[m.sup.2] at both sites (Table 3). Because of the shorter, steeper intertidal area at Shingle Spit, the total population was less than that of Fillongley, at 700 [+ or -] 500 individuals, compared with that at Fillongley of 2,000 [+ or -] 1,000 individuals.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Shell Assemblage Prey Preference
P. staminea was the most abundant of the drilled shells collected at Fillongley, followed by Macoma clams (Table 4). These were also the most common of the drilled shells collected at Shingle Spit, but the abundances were reversed, with Macoma spp. being the most abundant followed by P. staminea. The highest number of drilled shells were collected from the lowest tide stratum at both sites.
At Fillongley, 9 species were found with E. lewisii drill marks. The "other" group included M. arenaria, S. gigantea, P. tenuisculpta, C. nuttallii, and the wrinkled dogwinkle gastropod Nucella lamellosa. The diversity in the diet of E. lewisii was slightly lower at Shingle Spit, where 6 species were consumed. The "other" group was comprised of M. arenaria and S. gigantea.
When the proportions of the collected drilled shells were compared with the proportions of the species available in the community E. lewisii does not take clams in direct proportion to their availability (chi square = 1,739.74, P < 0.0001; Fig. 5). Even though V. philippinarum represented the species available in the highest proportion where the moon snails were found, the proportion of drilled shells collected for this species was very low.
When looking at each species individually, Ivlev's electivity coefficients (Ivlev 1961) showed that there was a preference for P. staminea, N. obscurata, Macoma spp., M. arenaria, S. gigantea, P. tenuisculpta, and C. nuttallii at Fillongley (Fig. 6D). V. philippinarum, R. retifera, L. californica, and T. carpenteri were avoided. At Shingle Spit, only P. staminea, Macoma spp., and S. gigantea were preferred, whereas all other species were avoided (chi square = 740.01, P < 0.0001; Fig. 7D). Differences in this feeding pattern were noted when each stratum was analyzed individually. P. staminea was a preferred prey item at both sites in every stratum with the exception of the low zone at Fillongley, where it was close to the no-preference line (Fig. 6A-C). S. gigantea was present only in the mid and low zones at both sites. Whenever it was present, it was a preferred prey species for E. lewisii. According to the Ivlev electivity coefficients (Ivlev 1961), Macoma clams were avoided at all tide heights at Shingle Spit, even though they were the most commonly collected drilled shell at that site (Fig. 7A-C). However, E. lewisii did show a preference for them when the study area was looked at as a whole. They were preferred prey in the mid and low zones at Fillongley.
Impacts of E. lewisii Predation on Intertidal Clan Communities
There were close to 3 million clams available in an area of 10,620 [m.sup.2] of beach at Fillongley within the range of E. lewisii. There was, on average, 228 clams/[m.sup.2]. At an overall density of 0.22 snails/[m.sup.2] in this area, E. lewisii, feeding at a rate of 0.09 clams/day, in 1 mo approximately 5,400 clams would be consumed (Fig. 8). This is 0.22% of the clam population in the study area. If these values are then converted to 6 mo and 1 y of feeding in the area, E. lewisii consumes 1.35% and 2.70% of the clam population, respectively. This would cost $5,913/y in sales at Fillongley, plus the cost inputs of time and applying antipredator netting and manually removing E. lewisii. It must be said that this cost would be if E. lewisii consumed only V. philippinarum for the year, but the rate used is based on a variety of prey species. The year values should also be considered high estimates because E. lewisii decreases its feeding rate during the winter months and densities decrease (Peitso 1980, Huebner & Edwards 1981, Peitso et al. 1994).
The impacts were similar at Shingle Spit (Fig. 8). There were fewer clams total at Shingle Spit (close to 800,000 clams) and 228 clams/[m.sup.2]. At the rate previously mentioned and a density of E. lewisii of 0.22 snails/[m.sup.2], 1,890 clams are consumed in 1 mo, which is 0.24% of the total clam population. Continuing at these feeding rates, E. lewisii consumes 1.47$ and 2.94% of the clam population in 6 and 12 mo, respectively, and costs the industry 52,069.55 in sales.
The work described here found that V. philippinarum is avoided by E. lewisii, suggested by the results of both prey preference experiments and observed shell assemblages. The only other study conducted on E. lewisii prey preferences found that only 0.4% of the drilled shells collected were V. philippinarum, indicating that this species is not favored (Bernard 1967). P. staminea was the preferred prey of E. lewisii based on our experiments. Bernard (1967), Harbo (2001), Peitso (1980), and Reid and Gustafson (1989) also found this preference. Beach shell assemblages also confirmed a preference for P. staminea. Despite the observed prey preference, the beach shell assemblages showed a diverse diet. At specific tide heights other prey were chosen, including Macoma spp., P. tenuisculpta, M. arenaria, S. gigantea, N. obscurata, and C. nuttallii. The beach shell assemblages gave important indications of the prey preferences of E. lewisii. The accuracy of this data is limited in that the drilled valves of thinner shelled prey species are not likely to persist as long in the habitat (Vignali & Galleni 1986).
[FIGURE 5 OMITTED]
Prey preference is common in naticid snails. Wiltse (1980a) found that Polinices duplicatus, an east coast naticid snail, ate 13 different species but showed preferences for M. arenaria and Gemma gemma. Euspira heros was shown to favor Macoma balthica and M. arenaria (Commito 1982). Spisula solidissima was preferentially consumed by E. heros (Weissberger 1999). Vignali and Galleni (1986) found that Donax trunculus was the species that was most attacked by the naticids from Piombino, Italy.
[FIGURE 6 OMITTED]
Preferences from the cage experiments may be attributed to the stratification of the 3 tested species within the sediment, because burial depth is a phenomenon that can affect prey preferences (Commito 1982). V. philippinarum lives very close to the sediment surface because of its short siphons (Meyer & Byers 2005). E. lewisii may burrow below V. philippinarum and therefore does not encounter it as readily as it does P. staminea and N. obscurata, species found deeper within the sediment.
The distribution of clams throughout the intertidal could also result in the preferences. V. philippinarum lives at the higher end of the range of E. lewisii, and therefore there is limited overlap in their distributions in the intertidal. However, this does not explain the observed preferences, because N. obscurata lives even higher on the intertidal than V. philippinarum and was consumed to a greater extent by E. lewisii.
Prey species is known to affect feeding rates (Moran 1985, Vignali & Galleni 1986, Rodrigues et al. 1987, Thiel et al. 2001). Our by-species feeding rates show that P. staminea was preyed upon at the highest rate by E. lewisii. N. obscurata, a newly introduced species in the area, was consumed at the second fastest feeding rate. V. philippinarum was the avoided prey type with the lowest feeding rate, even under monoculture type conditions. Bernard (1967) found that E. lewisii consumed P. staminea faster than it consumed S. gigantea and T. nuttallii, which supports the conclusions of this work.
The feeding rate of 0.09 clams per snail per day was determined for E. lewisii consuming a variety of available species. This is within the range found in previous studies (Peitso et al. 1994). An earlier study by Bernard (1967) found the feeding rate to be 0.25 clams per snail per day. However, in the study by Bernard (1967), snails were starved for 5 days prior to experimentation, placed in tanks with a limited amount of sediment, and all attempts and partially consumed clams were used in feeding rate calculations. Studies have shown that moon snails will not return to the same drill site to continue feeding on the prey item when they have been interrupted (Dietl & Alexander 1997, Kingsley-Smith et al. 2003). Thus, including drill attempts could have inflated the feeding rate. Peitso et al. (1994) found that the summer feeding rate was approximately 0.07 clams per snail per day, which is close to our 0.09 estimates. Our rate was determined to be 1 clam consumed every 14 days, a very slow feeding rate. Previous work on moon snail feeding rates has shown a wide range of feeding rates between snail species. E. heros had a maximum feeding rate of 1 clam per day (Weissberger 1999). P. pulchellus was found to consume 14.57 clams per snail per month at its maximum rate (Kingsley-Smith et al. 2003). Thus, feeding rates are not comparable between species.
[FIGURE 7 OMITTED]
Predator size, prey size, and temperature can all influence feeding rates. These factors must be considered when looking at feeding ecology because they can lead to an elevated feeding rate. Smaller snails have higher consumption rates (Edwards & Huebner 1977, Wiltse 1980b, Huebner & Edwards 1981, Peitso et al. 1994, Kingsley-Smith et al. 2003, Bea12006). Prey size can be optimized for best grip by the moon snail, which facilitates drilling and increases feeding rates (Wiltse 1980b, Commito 1982, Vignali & Galleni 1986). Prey can also reach a size refuge making them unavailable to predation (Rodrigues et al. 1987); however, we did not find any evidence to show that a size refuge was achieved at our study sites. Peitso (1980) found significant differences between the summer and winter feeding rate of E. lewisii, with the rate being highest during the summer. The rate determined in our study is a summer feeding rate over both years. The spring, fall, and winter rates are lower as a result of lower temperatures. Kingsley-Smith et al. (2003) and Weissberger (1999) found that moon snail feeding rates were dependent on temperature. Many naticid snails will actually stop feeding for 4 mo in the winter, as was seen in P. duplicatus (Huebner & Edwards 1981). This species stopped feeding completely at a temperature less than 5[degrees]C (Edwards & Huebner 1977). The density of E. lewisii in intertidal habitats decreases in the winter months, which would also lead to a decrease in impact (Peitso 1980). Therefore, the feeding rate determined in the current work is an upper limit, which must be considered when estimating the snail's impact on the community.
[FIGURE 8 OMITTED]
Other factors that can influence feeding rates include predator and prey density. Feeding rates in E. lewisii were not influenced by intraspecific density or by prey density, as was demonstrated by Peitso (1980). Although we did not test for this, we can assume that at higher prey densities on aquaculture leases, the feeding rates we determined would remain constant.
Prey preferences and the resulting feeding rates can be explained using the optimal foraging theory in which predators consume prey that leads to the highest energy gain in the least amount of time and energy input (Boggs et al. 1984). Naticid gastropod prey preference follows this hypothesis (Dietl & Alexander 1997). Savini and Occhipinti-Ambrogi (2006) found that moon snails maximize their energy intake by selecting a specific prey species that they can consume efficiently, rather than the immediately available species. E. lewisii followed this pattern, except where its preferred prey was not readily available. V. philippinarum is the numerically dominant species at both study sites, yet it was avoided in our experiments and the beach shell assemblages, where other species are available in lower numbers.
Feeding in E. lewisii is a large investment of energy, because they must spend quite a lot of time and energy drilling through the shell of its prey before feeding actually begins. Therefore, it needs to find prey that will facilitate these activities. Rodrigues et al. (1987) found that prey was selected based on a shell morphology that eases handling and reduces energy input. P. staminea's round and inflated shell morphology facilitates drilling at the umbo (Vignali & Galleni 1986, Reid & Gustafson 1989). Variations in shell thickness lead to variations in feeding rates and handling time. Minor changes in shell thickness can lead to dramatic changes in feeding rate. In a slow-feeding organism, such as E. lewisii, fractions of millimeters can increase drilling time by at least 25 h (Dietl & Alexander 1997). It may take longer to drill P. staminea because of its relatively thick shell, but it contains more calories than the other 2 species (Kirk 2007). Although N. obscurata has the lowest energy content, it may be selected over V. philippinarum because it has a thinner shell and takes less time to drill. In P. duplicatus, drilling alone took approximately 36 h on its preferred prey species, Mya arenaria (Boggs et al. 1984). Finding exact feeding rates in burrowing snails such as E. lewisii is complicated as a result of not being able to measure drilling times directly.
In eastern Canada and the United States, moon snail predation on commercially valuable shellfish has been considered to be high enough to warrant the use of public funds to control their populations. Commito (1982) proposed that moon snails are responsible for 96.5% of the mortality of M. arenaria. Predation is stated to be the most important factor determining juvenile clam survival in Maine, where 77% of clam mortality is attributed to the moon snail E. heros (Beal 2006). In BC, the code of practice put out by the British Columbia Shellfish Growers Association (2002) listed E. lewisii as one of several species that can have significant economic impact to the V. philippinarum industry. To protect their crop, the shellfish growers are removing E. lewisii from the intertidal.
Recent work has shown that feeding rates and impacts may be exaggerated. Clam deaths by crabs and other predators have been attributed to moon snails in some studies, implying that moon snail predation was over emphasized (Beal et al. 2001). Green (1968) estimated annual mortality rates of 28.2% from skate predation and other shell-destroying causes, 14.3% from crowding-related causes, and only 4% from naticid predation and this was by 2 different species. Predation by P. duplicatus was found to be only a minor source of mortality for G. gemma, one of its preferred prey species (Wiltse 1980b). Miron et al. (1985) found that the naticid E. heros was the predator that had the lowest feeding rate on all clam species tested compared with 2 sea star predators in eastern North America. Feeding rates in P. duplicatus were found to be less than previously believed (Huebner & Edwards 1981). Our work and the work performed by Peitso (1980) and Peitso et al. (1994), demonstrated that the feeding rates of E. lewisii are much lower than Bernard (1967) originally found. Our findings, as with those of Peitso et al. (1994) suggest that, over a year, about 3% of clam population mortality is the result of E. lewisii predation. This study stresses the importance of understanding the feeding ecology of a predator before suggesting antipredation measures.
E. lewisii's avoidance of V. philippinarum, low feeding rate, and low impact on the bivalve community can be applied to sustainable shellfish aquaculture practices. The results demonstrate that there is no longer a need to remove E. lewisii from intertidal lease areas, saving the time and energy of shellfish growers. The economic impact is limited, especially when the cost of antipredator netting and manual removal is considered. The impact to the intertidal ecosystem by aquaculture activities is thereby reduced and E. lewisii can be left in place to fulfill its ecological function.
Our study and the results of recent studies can lead to the general conclusion that moon snails have very low impacts on natural clam populations through predation activities. Biases with regard to the amount of moon snail prey could stem from the incriminating artifacts that are left behind, such as the bored shell, the number of which will accumulate over time, given a false impression of the numbers of clams actually preyed upon during a given period. Studies prior to 1990 have also been conducted under artificial conditions over short time periods, which lead to predation overestimates.
Funding for this work was provided by an Natural Sciences and Engineering Research Council of Canada strategic grant to L. Bendell-Young. Many thanks go out to Tracey L'Esperance for all her assistance in the field. Thanks also to Carolyn Allen, Chris Kowalchuk, Bruno L'Esperance, Jonathan Whiteley, and Wayne Kowalchuk for their support and assistance on various aspects of this research. Appreciation also goes out to Jenna Thomson, Mike White, and Charlotte Voss for taking such a keen interest in the project and helping out with data collection. Mike Hart provided constructive and helpful comments throughout this research.
Beal, B. F. 2006. Biotic and abiotic factors influencing growth and survival of wild and cultured individuals of the softshell clam (Mya arenaria L.) in eastern Maine. J. Shellfish Res. 25(2):461-474.
Beal, 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 juveniles of the soft-shell clam, Mya arenaria L., in Maine, USA. J. Exp. Mar. Biol. Ecol. 264(2):133-169.
Bernard, F. R. 1967. Studies on the biology of the naticid clam drill Polinices lewisii (Gould) (Gastropoda Prosobranchia). Nanaimo, B.C., Canada, Fisheries Research Board of Canada, 41 pp.
Boggs, C. H., J. A. Rice, J. A. Kitchell & J. F. Kitchell. 1984. Predation at a snail's pace: what's time to a gastropod? Oecologia 62(1):13-17.
British Columbia Ministry of the Environment, Oceans and Marine Fisheries Branch. 2008. Shellfish aquaculture in British Columbia. http://www.env.gov.bc.ca/omfd/fishstats/aqua/shellfish.html.
British Columbia Shellfish Growers Association. 2002. British Columbia shellfish farming industry: environment management system code of practice, http://bcsga.netfirms.com/wp-content/uploads/ 2007/08/enviro-mgmt-code-of-practice_02feb7.pdf.
Commito, J. A. 1982. Effects of Lunatia heros predation on the population dynamics of Mya arenaria and Macoma balthiea in Maine, USA. Mar. Biol. 69(2):187-193.
DeGraaf, J. D. & M. C. Tyrrell. 2004. Comparison of the feeding rates of two introduced crab species, Carcinus maenas and Hemigrapsus sanguineus, on the blue mussel, Mytilus edulis. Northeast. Nat. 11(2):163-166.
Department of Fisheries and Oceans. 2006. Economic potential of sea ranching and enhancement of selected shellfish species in Canada. http://www.dfo-mpo.gc.ca/Aquaculture/ref/SeaRanching_e.htm# AnnB.
Dietl, G. P. & R. R. Alexander. 1997. Predator--prey interactions between the naticids Euspira heros Say and Neverita duplicata Say and the Atlantic surfclam Spisula solidissima Dillwyn from Long Island to Delaware. J. Shellfish Res. 16(2):413-422.
Dudas, S. E., I. J. McGaw & J. F. Dower. 2005. Selective crab predation on native and introduced bivalves in British Columbia. J. Exp. Mar. Biol. Ecol. 325(1):8-17.
Edwards, D. C. & J. D. Huebner. 1977. Feeding and growth rates of Polinices duplieatus preying on Mya arenaria at Barnstable Harbor, Massachusetts. Ecology 58(6): 1218-1236.
Green, R. H. 1968. Mortality and stability in a low diversity subtropical intertidal community. Ecology 49(5):848-854.
Harbo RM. 2001. Shells and shellfish of the Pacific Northwest. Madeira Park: Harbour Publishing.
Huebner, J. D. & D. C. Edwards. 1981. Energy budget of the predatory marine gastropod Polinices duplicatus. Mar. Biol. (Berlin) 61(2-3): 221-226.
Ivlev, V. S. 1961. Experimental ecology of the feeding fishes. Scott D, translator. New Haven, CT: Yale University Press. 302 pp.
Kingsley-Smith, P. R., C. A. Richardson & R. Seed. 2003. Stereotypic and size-selective predation in Polinices pulchellus (Gastropoda: Naticidae) Risso 1826. J. Exp. Mar. Biol. Ecol. 295(2):173-190.
Kirk, M. 2007. Movement and foraging behaviours of surf scoters wintering in habitats modified by shellfish aquaculture. MS thesis. Simon Fraser University, Burnaby, BC.
Krebs, C. J. 1999. Manly's [alpha]. In: Ecological methodology, 2nd edition. Menlo Park, CA: Addison-Wesley Educational Publishers. pp. 483-486.
Krebs, C.J. 2001. Ecology: the experimental analysis of distribution and abundance, 5th edition. San Francisco: Benjamin Cummings.
Manly, B. F. J. 1974. Model for certain types of selection experiments. Biometrics 30(2):281-294.
Manly, B. F. J., P. Miller & L. M. Cook. 1972. Analysis of a selective predation experiment. Am. Nat. 106(952):719-736.
Meyer, J. J. & J. E. Byers. 2005. As good as dead? Sublethal predation facilitates lethal predation on an intertidal clam. Ecol. Lett. 8(2): 160-166.
Miron, G., D. Audet, T. Landry & M. Moriyasu. 2005. Predation potential of the invasive green crab (Carcinus maenas) and other common predators on commercial bivalve species found on Prince Edward Island. J. Shellfish Res. 24(2):579-586.
Moran, M. J. 1985. Effects of prey density, prey size and predator size on rates of feeding by an intertidal predatory gastropod Morula marginalba Blainville (Muricidae), on several species of prey. J. Exp. Mar. Biol. Ecol. 90(2):97-105.
Peitso, E. 1980. Predation by the moon snail, Poliniees lewisii (Gould), on the littleneck clam, Protothaca staminea (Conrad). MS thesis. Simon Fraser University, Burnaby, BC.
Peitso, E., E. Hui, B. Hartwick & N. Bourne. 1994. Predation by the naticid gastropod Poliniees lewisii (Gould) on littleneck clams Protothaca staminea (Conrad) in British Columbia. Can. J. Zool. 72(2):319-325.
Reid, R. G. B. & B. D. Gustafson. 1989. Update on feeding and digestion in the moon snail Polinices lewisii (Gould, 1847). Veliger 32(3):327.
Rodrigues, C. L., S. Nojima & T. Kikuchi. 1987. Mechanics of prey size preference in the gastropod Neverita didyma preying on the bivalve Ruditapes philippinarum. Mar. Ecol. Prog. Set. 40(1-2):87-93.
Savini, D. & A. Occhipinti-Ambrogi. 2006. Consumption rates and prey preference of the invasive gastropod Rapana venosa in the northern Adriatic Sea. Helgol. Mar. Res. 60(2):153-159.
Schwarz, C.J. 2005. Stat 403/Stat 650: intermediate sampling and experimental design and analysis: course notes. Simon Fraser University, Burnaby, BC.
Thiel, M., N. Ullrich & N. Vasquez. 2001. Predation rates of nemertean predators: the case of a rocky shore hoplonemertean feeding on amphipods. Hydrobiologia 456:45-57.
Vignali, R. & L. Galleni. 1986. Naticid predation on soft bottom bivalves: a study on a beach shell assemblage. Oebalia (Taranto) 13:157-177.
Weissberger, E. J. 1999. Additive interactions between the moon snail Euspira heros and the sea star Asterias forbesi, two predators of the surfclam Spisula solidissima. Oeeologia 119(3):461-466.
Whiteley, J. 2005. Macroinvertebrate community responses to clam aquaculture practices in British Columbia, Canada. MS Thesis. Simon Fraser University, Burnaby, BC.
Whiteley, J. A. & L. I. Bendell-Young. 2007. Ecological implications of intertidal mariculture: observed differences in bivalve community structure between farm and reference sites. J. Appl. Ecol. 44(3):495-505.
Wiltse, W. I. 1980a. Effects of Poliniees duplicatus (Gastropoda: Naticidae) on infaunal community structure at Barnstable Harbor, Massachusetts, USA. Mar. Biol. (Berlin) 56(4):301-310.
Wiltse, W. I. 1980b. Predation by juvenile Polinices duplicatus (Say) on Gemma gemma (Totten). J. Exp. Mar. Biol. Ecol. 42(2):187-199.
NICOLA COOK* AND LEAH BENDELL-YOUNG
Department of Biological Sciences, Simon Fraser University, 8888 University Drive, Burnaby, BC, Canada V5A 1S6
* Corresponding author: E-mail: email@example.com
TABLE 1. Length of the 3 tide strata at each site. Tide Height Zone Length (m) Tide (m above chart Zone datum) Fillongley Shingle Spit High 2.3-1.7 30 25 Mid 1.7-1.3 67 12 Low 1.3-0.7 80 20 TABLE 2. Total clam abundance by species at Fillongley and Shingle Spit for each stratum [+ or -] 95% confidence interval. Site Fillongley High Tide Mid Tide Species Zone Zone Protothaca 190,000 [+ or -] 40,000 600,000 [+ or -] 200,000 staminea Venerupis 540,000 [+ or -] 80,000 600,000 [+ or -] 50,0000 philippinarum Nuttallia 300 [+ or -] 600 1,000 [+ or -] 3,000 obscurata Macoma spp. 8,000 [+ or -] 8,000 40,000 [+ or -] 20,000 Other 2,100 [+ or -] 2,000 20,000 [+ or -] 10,000 Total 750,000 [+ or -] 80,000 1,200,000 [+ or -] 600,000 Euspira lewisii 0 0 Site Fillongley Shingle Spit Low Tide High Tide Species Zone Zone Protothaca 360,000 [+ or -] 60,000 70,000 [+ or -] 20,000 staminea Venerupis 0 290,000 [+ or -] 80,000 philippinarum Nuttallia 0 2,0000 [+ or -] 2,0000 obscurata Macoma spp. 60,000 [+ or -] 30,000 10,000 [+ or -] 10,000 Other 70,000 [+ or -] 20,000 20,000 [+ or -] 10,000 Total 500,000 [+ or -] 70,000 410,000 [+ or -] 60,000 Euspira lewisii 2,000 [+ or -] 3,000 0 Site Shingle Spit Mid Tide Low Tide Species Zone Zone Protothaca 39,000 [+ or -] 9,000 30,000 [+ or -] 10,000 staminea Venerupis 20,960 [+ or -] 20,000 2,000 [+ or -] 3,000 philippinarum Nuttallia 0 0 obscurata Macoma spp. 90,000 [+ or -] 30,000 180,000 [+ or -] 30,000 Other 8,000 [+ or -] 4,000 6,000 [+ or -] 3,000 Total 160,000 [+ or -] 2,000 220,000 [+ or -] 40,000 Euspira lewisii 200 [+ or -] 300 300 [+ or -] 500 Fillongley, n = 72; Shingle Spit n = 54. TABLE 3. Density of E. lewisii at Fillongley and Shingle Spit (measured in density per square meter [+ or -] 95% confidence interval) and total abundance in the survey area [+ or -] 95% confidence interval. Site Variable Fillongley Shingle Spit Density (no./[m.sup.2]) 0.2 [+ or -] 0.2 0.2 [+ or -] 0.1 Total population 2,000 [+ or -] 1,000 700 [+ or -] 500 Fillongley, n = 50; Shingle Spit, n = 50. TABLE 4. Raw numbers of drilled shells collected in each stratum at each site with totals. Fillongley Species High Mid Low Total Protothaca staminea 16 278 623 917 Venerupis philippinarum 1 6 2 9 Nuttallia obscurata 0 0 0 0 Macoma spp. 0 67 176 243 Other 1 38 103 142 Total 18 389 904 1,311 Shingle Spit Species High Mid Low Total Protothaca staminea 12 106 157 275 Venerupis philippinarum 1 4 2 7 Nuttallia obscurata 1 0 0 1 Macoma spp. 0 51 334 385 Other 3 22 26 51 Total 17 183 519 719