Available online at www.nznaturalsciences.org.nz

Same enemy, same response: predator avoidance by an invasive and native snail

Joseph R. Holomuzki1,3 & Barry J. F. Biggs2

1Department of Evolution, Ecology, and Organismal Biology, The Ohio State University, Mansfield, Ohio, U.S.A. 2National Institute of Water and Atmospheric Research Limited, Christch- urch, New Zealand 3Corresponding author’s email: [email protected]

(Received 21 December 2011, revised and accepted 25 April 2012)

Abstract

Novel or highly effective antipredator traits can facilitate successful invasions by prey . Previous studies have documented that both the native New Zealand snail, antipodarum, and the highly successful, worldwide invader Physella (Physa) acuta, seek benthic cover to avoid fish predators. We asked whether an invasive, South Island population of Physella has maintained this avoidance response, and if so, how it compared to that of Potamopyrgus that has coevolved with native fish. We compared patterns of sediment surface and subsurface use between snail species and between sizes of conspecifics both in a 2nd –order reach with predatory fish and in a laboratory experiment where fish presence (common bullies [Gobiomorphus cotidianus]) was manipulated. Both snails sought protective sediment subsurfaces when with bullies. Proportionally, sediment subsurface use both in field collections and in fish treatments in the experiment were similar between species. Field-captured bullies ate more Physella than Potamoprygus, suggesting different consumptive risks, but overall, few snails were consumed. Instream densities and sizes (shell length) of Physella were greatest on cobbles, where most (>90%) egg masses were found. Potamopyrgus densities were evenly distributed across sand, gravel, and cobbles. Shell lengths were similar between snails on sediment subsurfaces and surfaces during daylight for both species, suggesting this avoidance response was not size-dependent. For Physella, retaining the ability to seek benthic refugia when exposed to novel predators likely greatly contributes to its invasion success. Given both snails are dominant primary consumers in many freshwater ecosystems, surface-density increases resulting from predation pressure decreases would predictably significantly affect system trophic dynamics.

Key words: bully predation – – Potamopyrgus antipodarum – sediment surface and subsurface use – shell length – sediment size

New Zealand Natural Sciences (2012) 37: 11-24. © New Zealand Natural Sciences 12 New Zealand Natural Sciences 37 (2012)

Introduction hence displace other macroinvertebrates and fish (Lowe & Hunter 1988; Win- Novel or highly effective antipredator terbourn & Fegley 1989; Biggs & Lowe traits can facilitate successful invasions 1994). In New Zealand, their popula- by prey species. Traits that allow benthic tions seldom reach high densities, in part macroinvertebrate invaders to effectively because of predation by native and non- deter or avoid native predators, and hence native fishes (Levri 1998; Holomuzki & to become successful, include armouring Biggs 2006; Holomuzki 2010). (spines) (Zaranko et al. 1997; Levri et Predation risk might be a key driver al. 2007), inedibility (Strayer 2009), of microspatial patterns of both snails. activity reductions (Pennuto & Keppler Potamopyrgus generally resides on a 2008), and movement to spatial refugia broad range of sediment sizes (Jowett et (Dalesman et al. 2007). Identifying and al. 1991), but in streams, is most dense assessing the relative effectiveness of on cobbles in slow current speeds (≤ antipredator traits of native and non- 15 cm/s) (Holomuzki & Biggs 2007). native prey is important to predicting Regardless of sediment size, in streams the potential rate and geographical extent with predatory fish, ~80% of all Pota- of invader spread and impacts on native mopyrgus occur in sediment subsurfaces species numbers and assemblages (Levine with limited predator access (Holomuzki et al. 2003; Temperton et al. 2004). et al. 2009). Moreover, Potamopyrgus This study compares predator avoid- will move to protective sediment sub- ance behaviours by a native and non- surfaces when in the presence of mol- native snail. Potamopyrgus antipodarum luscivorous bullies (Gobiomorphus spp.), (Gray 1843) is native to New Zealand a widespread native fish in New Zealand streams and lakes, where it commonly streams and lakes (Holomuzki & Biggs co-exists with the highly successful, 2006). Similarly, Physella acuta will worldwide invader Physella (Physa) acuta move under cover when in the physical (Draparanaud 1805) (McCarter 1986; or chemical presence of pumpkinseed Cope & Winterbourn 2004; Kristensen sunfish (Lepomis gibbosus) (Turner 1996; & Closs 2008), apparently introduced Turner & Montgomery 2003) and creek from the Mediterranean region of Europe chub (Semotilus atromaculatus) (Turner or North America ca. 40 years ago (Win- et al. 2000), important molluscivores in terbourn 1973). Potamopyrgus is now also eastern North America (Brönmark et al. found in southeastern Australia (Sch- 1992). However, for invasive populations reiber et al. 1998; 2003), Japan (Shimada in New Zealand, it is unclear whether & Urabe 2003), Europe (Statzner 1981), Physella has retained this response, and the Laurentian (Zaranko et how this response, if retained, compares al. 1997), and the western to predator-induced microspatial shifts (Dybdahl & Kane 2005). So, both snails by Potamopyrgus that has coevolved with also widely co-occur as non-native spe- native predators. cies. When at high densities or biomass, The magnitude of any avoidance re- by both snails can significantly sponse might also depend on predation reduce algal biomass and richness, alter risk and prey vulnerability. Bullies eat algal physiognomy and composition, and both Physella and Potamopyrgus (Sagar Holomuzki & Biggs: Predator avoidance by invasive and native snails 13

& Glova 1995; Wilhelm et al. 2007), snail size distributions. but it is unknown whether consumptive risk from bullies differs between these Methods snails. Non-native, young brown trout (Salmo trutta) in New Zealand streams Snail and sediment field sampling apparently eat similar numbers of Pota- mopyrgus and Physella (Holomuzki 2010), Snails were sampled in the East Branch of even though Physella is larger-bodied the Kaiapoi River (a.k.a. Silverstream; lat (~7 mm average length), softer-shelled, 43.25°S, long 172.35°E) 26 km north of and energetically more profitable than Christchurch on 1 and 2 February 2011. Potamopyrgus (4–6 mm average length) Snails were collected in a ~25-m-long, (McCarter 1986). Moreover, in North ~11-m-wide reach (current velocity: America, Physa integra (= P. acuta, Dillon 5–16 cm/s; water depths: 12–31 cm) et al. 2002) are smaller at sites with fish containing mainly coarse gravel to small than without fish (Bernot & Whittinghill cobbles armouring sand. Macrophytes, 2003), but it is unclear if snails are smaller mostly watercress (Nasturtium officinale), because of size-selective predation or be- were present only along shorelines. The cause larger snails exhibit stronger avoid- predominant fishes in this 2nd-order ance behaviours that reduce predation reach were common bullies, short-finned risk than smaller ones. For Potamopyrgus, eels (Anguilla australis), and brown trout small, non-parasitized individuals (<3.8 (Holomuzki & Biggs, 2006). Indigenous mm length) are more active than large Paranephrops zealandicus crayfish were individuals on tops of rocks during day- absent (JRH, pers. obs.). light (Levri & Lively 1996), suggesting The reach was mapped to create a grid predation risk might vary with body size. system for sampling snails and sediments. Here, we present results from field The grid contained 2 rows, each 1-m wide collections, gut-evacuation observations, and 5 m-long near mid-channel, and and a laboratory experiment to examine spaced 1-m apart. Each row was divided 1) how densities of each snail vary with into twenty 0.5 x 0.5 m plots. Plots were sediment size, 2) how patterns of sedi- used to collect snails and sediments, and ment subsurface use differ between Phy- the 1-m wide space between the rows sella and Potamopyrgus and between sizes in the middle of the reach served as a of conspecifics when with predatory fish, walkway to minimize disturbance in and 3) whether these snails are equally sample plots. Twenty one of the 40 plots important prey items for common bul- within the grid were randomly selected lies (Gobiomorphus cotidianus McDow- for sampling with a cylindrical stovepipe all). Microdistributional patterns of P. core sampler (40 cm height, 25.4 cm di- acuta in streams are unknown, but clearly ameter). The sampler was pushed ~5 to 15 important to understanding trophic cm into sediments and all snails and sedi- relationships. We also relate egg mass dis- ments within the sampler were collected. tributions to sediment sizes for Physella, However, collection procedure varied but not for ovoviviparous Potamopyrgus, with water depth at plots. In depths <25 because egg-laying habitat preferences cm, all epibenthic snails in the sampler might affect sediment use patterns and were removed by suctioning, then sedi- 14 New Zealand Natural Sciences 37 (2012) ments were removed with a dip net (0.5 x 10 in the laboratory between north- and 0.5-mm mesh) to a depth of ~5 to 15 cm. west-facing windows to expose fish to a In depths from ~25 to 31 cm, epibenthic natural photoperiod (14–15 h daylight). snails were collected by removing the Metal screens (2 x 2-mm mesh) atop tubs sampler and suctioning snails within the retained fish. Tubs were substrate-free to border ring outlined in the sediments. A facilitate recovery of egested snails, and view box was sometimes used to detect fish were not fed while in tubs. Snails epibenthic snails within the border. After egested by each fish were siphoned ~19 h suctioning, the sampler was repositioned and ~42 h after fish were placed in tubs. on the border ring, and sediments were Snails retrieved from each time period removed as described earlier. Sediments were identified, counted, and measured were gently placed in a white plastic tray (shell length). All fish were released (50-cm diameter), and endobenthic snails at their site of collection after the last were hand-picked from sediments. Epi- siphoning. and endobenthic snails from each sample Laboratory experiment were placed in separate 100-mL plastic jars and preserved in 80% ethanol. In Bullies were captured by electrofishing on addition, Physella egg masses attached to 11 February 2011 from the same reach stones were counted in each sample. The electrofished on 2 February. Immediately predominant sediment type (sand, gravel after capture, fish were individually [4–32 mm, a-axis], or pebble to small placed in 17.0 x 17.0 x 8.5 cm high plastic cobbles [32–128 mm]; Wentworth scale) tubs each filled with 1.5 l of high-quality in each sample was determined visually, aquifer water (19° ± 1°C; O2: ~8.0 mg/l). and only one sample was taken per plot. Fish were kept indoors and starved for 27 Counts and identifications of snails were h before moving them to experimental made with a dissecting microscope in the tubs. laboratory, and shell lengths (mm, outer On 12 February, 20 white plastic tubs lip of to apex) were measured (size as above) were each filled with a with an ocular micrometer. ~3–4-cm thick layer (~1500 g) of pebble and small cobbles (Wentworth scale; par- Snail consumption by bullies ticle size, a-axis: 50.2 ± 0.2 mm [mean ± Twenty common bullies were electrofished SE], 33–84 mm [range], n = 40) collected on 2 February (1400–1500 h) from the from riparian areas of the East Branch. East Branch of the Kaiapoi in a ~15-m Sediments were thoroughly washed to long stretch ~25 m downstream of the remove periphytic . High-quality snail collection site. Immediately after aquifer water (1 l) was added to each tub, capture, fish were placed in white plastic so fishless treatments also were devoid of tubs (17.0 x 17.0 x 8.5 cm high). Tubs fish chemical cues. Tubs with sediments were filled with 1.5 l of water from the were placed in a 2 x 10 array between East Branch, maintained at 18°C (± 1°C), north- and west-facing windows to ex- and contained one fish. Water in tubs pose organisms to a natural photoperiod. was not aerated but was changed after One bully (mean ± SE: 4.62 ± 0.6 cm ~20 h to maintain oxygen levels (8.0 total length) captured the day before was mg/l). Tubs were aligned in 2 rows of placed in each of 10 randomly selected Holomuzki & Biggs: Predator avoidance by invasive and native snails 15 tubs 24 h before adding snails (n = 10 Statistical analyses replicates both for fish and fishless treat- For field collections, a one-way analysis ments). Black plastic screens (7 x 7-mm of variance (ANOVA) was used to mesh) placed atop tubs, even fishless ones, determine whether overall channel retained bullies. densities (log[x + 1]-transformed) differed On 13 February, ten similarly snails of between snails (SYSTAT v9.0). Two-way each species (Physella: 2.9–7.0 mm shell ANOVAs were used to test whether length [range]; Potamopyrgus: 3.0–6.1 Physella and Potamopyrgus densities mm, all smooth-shelled [body whorls (log[x + 1]-transformed) and sizes (shell not ridged and devoid of spines; Haase length; log(x)-transformed) differed 2003]) collected from the East Branch among sediment types and between were added to each tub between 1000 sediment surfaces and subsurfaces. Mean and 1330 h. These densities (~350 snails surface and subsurface sizes of snails in m-2) are representative of field densities each sample were used in the ANOVA. for Potamopyrgus, but are high for Physella Significant sediment-density or -size (Holomuzki & Biggs 2006; Holomuzki effects were followed by Bonferroni 2010). We chose to use equal densities pairwise comparisons (alpha = 0.05). and similar sizes of snails to equalize In the laboratory experiment, a 2-way consumptive risk. Twenty four hours after ANOVA was used to test whether bully additions, screens were removed, and all presence differentially affected subsurface snails visible on surface sediments and use (arcsine[x]-transformed proportions) tub walls, and floating, were suctioned by each snail species. Separate two-way from tubs. Although these snails were ANOVAs were also used to determine collectively considered as epibenthic, we whether bully presence affected mean separately noted the number and species snail size (log[x]-transformed shell of those floating to assess species-specific lengths) on cobble surfaces and subsur- differences in this possible predator- faces for each snail. A G-test of independ- induced dispersal mechanism (Crowl ence with Williams’s correction tested & Covich 1990). After these removals, whether the number of floating snails in sediments from each tub were thoroughly different predator environments (fish, hand-sorted to retrieve all snails from no fish) was independent of snail species subsurface sediments. Snails of each spe- (Sokal & Rohlf 1995). cies recovered from surface and subsur- face sediments were counted, preserved in 80% ethanol, and measured (shell Results length). We calculated proportions of snails in subsurface sediments by dividing numbers found in subsurface sediments Instream sediment use by the total number retrieved to account Potamopyrgus was significantly denser for fish consumption. All fish were -re than Physella in stream-channel sediments leased in the East Branch at the end of (mean ± SE, Potamopyrgus: 395 ± 54 nos. the experiment. m-2; Physella: 29 ± 7 nos. m-2) (1-way

ANOVA: F1,36 = 55.218, P < 0.001). 16 New Zealand Natural Sciences 37 (2012)

Densities varied across sediment types for egested prey item was amphipods. Physella (2-way ANOVA: F = 5.360, 2, 36 Laboratory experiment P = 0.009), but not for Potamopyrgus

(F2,36 = 0.763, P = 0.474) (Figure 1A, Both Physella and Potamopyrgus responded 1B). Physella was denser on cobbles than to bully presence by increasing their use of on sand (Bonferroni comparison: P = cobble subsurfaces (2-way ANOVA, fish

0.010), but not between sand and gravel effect:F 1,36 = 12.357, P < 0.001) (Figure (P = 0.326) and cobbles and gravel (P = 3). However, this response did not differ

0.296). Coincidently, 22 of 24 Physella between species (snail effect:F 1,36 = 2.000, egg masses were found on cobbles; the P = 0.166). Although very few snails were other two in gravel. found floating in tubs (Physella: 0, 3; Densities of both snails were greater Potamopyrgus: 6, 3; with and without fish, on sediment subsurfaces than on surfaces respectively), the behaviour was species-

(2-way ANOVAs, Physella: F1,36 = 3.795, dependent (G = 5.182, df = 1, P < 0.025).

P = 0.054; Potamopyrgus: F1,36 = 4.261, No difference in Physella shell length P = 0.045) (Figure 1A, 1B). However, was detected between surface and subsur- subsurface densities of both species were face sediments in tubs with (grand means independent of sediment type, based on ± 1SE; surface vs. subsurface: 4.92 ± 0.26 non-significant interaction terms (bothP vs. 4.58 ± 0.11 mm) and without fish > 0.15). No Physella were found on sand (4.55 ± 0.2 vs. 4.62 ± 0.2 mm) (surface surfaces (Figure 1A). effect: F1,35 = 0.496, P = 0.486; fish ef-

Physella shell length differed among fect: F1,35 = 0.739, P = 0.396). Likewise, sediment types (F2,13= 4.868, P = 0.026), Potamopyrgus size did not differ between with snails being larger on cobbles than surface and subsurface sediments in fish on gravel (Bonferroni comparison, P (surface vs. subsurface: 4.03 ± 0.16 vs. = 0.049; Figure 2A). In contrast, Pota- 4.08 ± 0.11 mm) and fishless tubs (4.31 mopyrgus shell length did not differ ± 0.1 vs. 4.20 ± 0.11 mm) (surface effect: among sediment types (F2,33 = 0.715, F1,36 = 0.082, P = 0.776; fish effect:F 1,36 = P = 0.496) (Figure 2B). For both spe- 2.556, P = 0.119). cies, shell lengths were similar between All snails were recovered from fish and individuals on sediment subsurfaces fishless tubs, except for one Physella in a and surfaces (Physella: F1,13 = 0.341, P = fish treatment.

0.569; Potamopyrgus: F1,33 = 3.661, P = 0.064). Discussion Snail consumption by bullies Our results suggest that invasive Physella A total of 7 snails, 6 Physella (mean ± has retained the ability to seek benthic SE; shell length: 2.1 ± 0.3 mm) and 1 cover when exposed to a native New Potamopyrgus (1.5 mm), were eaten by 5 Zealand predator in a novel environment. of 20 bullies. Four fish egested a total of Moreover, the strength of the response is 5 snails after ~19 h in tubs, while one fish similar to that of native Potamopyrgus, egested 2 Physella after ~42 h. All egested based on densities and proportions of snails were dead. The most commonly sediment subsurface use by both snails Holomuzki & Biggs: Predator avoidance by invasive and native snails 17

Surface A. Physella 40 Subsurface ab b 30 5

20

1 a 10 ) 2

0

600 B. Potamopyrgus a 500

Density (nos./m Density 400 a 5 300 7 7 a

200

100

0 Sand Gravel Cobble

Figure 1. Mean (+1 SE) densities of Physella acuta (A) and Potamopyrgus antipodarum (B) on surfaces and subsurfaces in sediment types in the East Branch of the Kaiapoi River (n = 5, 8, and 8 samples from sand, gravel, and cobbles, respectively). Different lowercase letters represent significant differences in densities between sediment types (Bonferroni comparisons,alpha = 0.05). Densities of both snails were greater on sediment subsurfaces than on surfaces (both P ≤ 0.054). 18 New Zealand Natural Sciences 37 (2012)

A. Physella Surface 6 ab Subsurface b 5 1

a 5 4 6 5 3

1 2

1

0

5 B. Potamopyrgus

a 4 a a

Shell length (mm) 8 5 3 5 7 7 7 2

1

0 Sand Gravel Cobble

Figure 2. Mean (+1 SE) shell lengths of Physella acuta (A) and Potamopyrgus antipodarum (B) on surfaces and subsurfaces in sediment types in the East Branch. Numbers in bars indicate the number of surface and subsurface samples in which the species was collected (total number of snails measured: n = 32 and 420 for Physella and Potamopyrgus, respectively). Different lowercase letters represent significant differences in shell length between sediment types (Bonferroni comparisons, alpha = 0.05). Shell lengths of both snails were similar between sediment surfaces and subsurfaces (both P ≥ 0.064). Holomuzki & Biggs: Predator avoidance by invasive and native snails 19

in field collections and the laboratory & Covich 1991 [Procambarus simulans]; experiment, respectively. This result is DeWitt et al. 1999 [Orconectes rusticus]; somewhat surprising, given our gut- McCarthy & Fisher 2000 [P. clarki]; egestion observations hint Physella Bernot & Turner 2001 [O. rusticus]), and might be preferred by Gobiomorphus, this response is stronger when crayfish size selective feeders on benthic are actively foraging or sifting through macroinvertebrates (Gregory et al. 2007). substratum interstices (McCarthy & However, this result might be expected Fisher 2000). We did not differentially if the risk of predation is similarly score snails at the waterline on tub walls perceived by both snails (Levri 1998) and in our fish and fishless treatments, so if subsurface use is a generalised, non- we cannot say whether Physella is more discriminatory response to fish predators, likely than Potamopyrgus to exhibit regardless of the risk they pose. Yet, this response in the presence of bullies. movement behaviours by Physella suggest However, we did count floating snails, they can identify some enemy types, and albeit counts were low, Potamopyrgus perhaps by processing different chemical appears more prone than Physella to float and physical cues (Turner et al. 1999). In to escape risky habitats. This behaviour contrast to their reactions to fish,Physella might also depend on the strength of will crawl to or above the waterline to injured-conspecific cues (McCarthy & temporarily avoid crayfish (Alexander Fisher 2000), which were negligible in

Figure 3. Mean (+1 SE) subsurface sediment use by Physella and Potamopyrgus in fish and fish treatments in the laboratory experiment (n = 10 replicates/treatment). P-value represents ANOVA results for a fish effect on subsurface use. 20 New Zealand Natural Sciences 37 (2012) our experiment, and might, like other appear that movements to endobenthic antipredator behaviours, be amplified surfaces by both species are independent in small-mesocosm experiments with of sediment size. standing water (Peckarsky et al. 2002). In Field samples show that Physella shell addition, both Physella and Potamopyrgus lengths differed among sediment sizes. are significantly more abundant on Physella were larger on cobbles, where structurally-complex macrophytes than most egg masses were detected, than on on -coated stones on the bottoms gravel. Larger, egg-bearing snails might of lakes and streams (Kelly & Hawes prefer larger, more stable sediments to 2005; Holomuzki 2010), but it is unclear attach egg masses. In addition, cob- whether plant beds are actively sought to bles provide a larger surface area for avoid predators. Thus, the use of less risky periphyton growth, and hence a larger habitats regardless of predator identity surface for foraging, than gravel (Brown appears to be a key general defence for 1982). We did not detect a relationship both snails that likely contributes to their between Potamopyrgus shell length and widespread invasion success. sediment size, nor a statistically signifi- Although spatial responses were simi- cant size difference P( = 0.064) between lar, temporal relationships between move- individuals on sediment subsurfaces and ment and predator presence or kind could sediment surfaces during sample times vary between snails. Our study design (~0900–1700 h). Moreover, we did not did not allow us to assess or compare the detect a body-size difference in space rapidity of the response by each species, use in the laboratory experiment, where nor the length of time individuals remain the size range of snails used for each in subsurface sediments after exposure to species (none < 2.9 mm shell length) fish, both of which could affect consump- was narrower than that in Silverstream, tion risk. Potamopyrgus can remain on which likely hindered the probability of rock undersides for >20 h when exposed detecting a difference. Levri & Lively only to the chemical cues of bullies in (1996) reported that small, trematode laboratory tanks (Levri 1998), and for ()-free Potamopyrgus were ~8 h (overnight) when coexisting with more active than large individuals on nocturnal predatory fish in natural tops of littoral rocks in Lake Alexandrina, streams (Holomuzki et al. 2009). Simi- South Island, during daylight (~1530 larly, Physella can remain in spatial refugia h). Although we did not assess parasit- (i.e., under ceramic tile, at or above the ism, epibenthic snails in our study were waterline) for at least 12 h to avoid fish probably also largely Microphallus-free, (Bernot & Whittinghill 2003). However, given frequency of trematode infection unlike Potamopyrgus, use of benthic cover of the Silverstream population is very by Physella does not apparently differ low (Holomuzki et al. 2009). Size differ- between day and night, at least when ences between Potamopyrgus on sediment exposed to diurnally-feeding creek chub surfaces and subsurfaces in Silverstream (Bernot & Whittinghill 2003). Whether might be more prominent at night when Physella responds to bullies faster than relatively few snails are active on rock Potamopyrgus, or remains in protective surfaces, apparently to reduce encoun- sediment subsurfaces longer than Pota- ters with nocturnally hunting bullies mopyrgus, is unknown. However, it does and shortfinned eels (Anguilla australis) Holomuzki & Biggs: Predator avoidance by invasive and native snails 21

(Holomuzki et al. 2009). Even so, it uncommon in sand-dominated habitats. is unclear whether trade-offs between These scenarios define perspectives for foraging return and predator avoidance future multipopulational studies that are associated with endobenthic activity. investigate anti-predator phenotypic Potamopyrgus and Physella are herbivore- reactions by these snails across various detritivores (Dillon 2000; Cope & Win- predator regimes and habitat types. terbourn 2004), capable of consuming diatom-dominated biofilms on stone sur- Acknowledgements faces (Talbot & Ward 1987; Suren 2005) or heterotrophic biofilms in sediment We thank Carole Holomuzki for help subsurfaces (Rounick & Winterbourn collecting and sorting snails, Denis 1983). So, energetic costs associated with Maindonald and Carole Holomuzki for endobenthic activity are likely minimal. help electrofishing bullies, Ben Devitt In addition, hydrodynamic forces (drag) of Salmon Smolt NZ Ltd. for access are typically much stronger near tops of to Silverstream. This work was partly rocks than in sediment subsurface voids supported by grant C01X0308 (Water (Dole-Olivier et al. 1997; Holomuzki Allocation: Protection of Instream Values) & Biggs 2000). Thus, snails residing in from the New Zealand Foundation sediment subsurfaces apparently gain pro- for Research Science and Technology tection from predators, access to a viable to BJFB. Helpful comments on the food source, and relief from drag forces. manuscript were provided by two referees. Comparative studies like the one de- scribed here are necessary to determine References whether success depends on different mechanisms in different Alexander, J.E. & Covich, A.P. (1991). communities. It appears that both Phy- Predation risk and avoidance behavior sella and Potamopyrgus have the ability in two freshwater snails. Biological to move from areas of high predation Bulletin 180: 387–393. risk to low predation risk in a variety of Bernot, R.J. & Turner, A.M. (2001). predator regimes, and that this ability is Predator identity and trait-mediated not amplified by invasive Physella. From indirect effects in a littoral food web. an applied perspective, this means that Oecologia 129: 139–146. surface-density increases of these domi- Bernot, R.J. & Whittinghill, K. (2003). nant primary consumers, resulting from Population differences in effects of fish natural or anthropogenic perturbations on Physa integra refuge use. American that reduce predator abundance, could Midland Naturalist 150: 51–57. indirectly affect periphyton biomass and Biggs, B.J.F. & Lowe, R.L. (1994). composition, the cycling of nutrients, Responses of two trophic levels to patch and food availability for higher trophic enrichment along a New Zealand level consumers (Lowe & Hunter 1988; stream continuum. New Zealand Winterbourn & Fegley 1989; Biggs & Journal of Marine and Freshwater Lowe 1994). This potential effect on Research 28: 119–134. trophic and benthic processes might be Brönmark, C., Klosiewski, S.P. & modified by bed-sediment composition, Stein, R.A. (1992). Indirect effects of particularly for Physella which seems predation in a freshwater, benthic food 22 New Zealand Natural Sciences 37 (2012)

chain. Ecology 73: 1662–1674. 86: 1592–1601. Brown, K.M. (1982). Resource overlap Gregory, M.P., Davis, A.R. & Ayre, D.J. and competition in pond snails: an (2007). Diet and feeding periodicity experimental analysis. Ecology 63: of Cox’s gudgeon Gobiomorphus coxii 412–422. (Krefft) in a south-eastern Australian Cope, N.J. & Winterbourn, M.J. (2004). stream. Journal of Fish Biology 71: Competitive interactions between 993–1006. two successful molluscan invaders of Haase, M. (2003). Clinal variation in freshwaters: an experimental study. shell morphology of the freshwater Aquatic Ecology 38: 83–91. gastropod Potamopyrgus antipodarum Crowl, T.A. & Covich, A.P. (1990). along two hill-country streams in New Predator-induced life-history shifts in a Zealand. Journal of the Royal Society of : a chemically mediated New Zealand 33: 549–560. and phenotypically plastic response. Holomuzki, J.R. (2010). Within- Science 247: 949–951. reach spatial variability of snails and Dalesman, S., Rundle, S.D. & Cotton, molluscivory by juvenile brown trout. P.A. (2007). Predator regime influences New Zealand Journal of Marine & innate anti-predator behaviour in Freshwater Research 44: 187–195. the freshwater gastropod Lymnaea Holomuzki, J.R. & Biggs, B.J.F. (2000). stagnalis. Freshwater Biology 52: 2134– Taxon-specific responses to high-flow 2140. disturbance in streams: implications DeWitt, T.J., Sih, A. & Hucko, J.A. for population persistence. Journal (1999). Trait compensation and of the North American Benthological cospecialization in a freshwater snail: Society 19: 670–679. size, shape and antipredator behaviour. Holomuzki, J.R. & Biggs, B.J.F. Behaviour 58: 397–407. (2006). Habitat-specific variation Dillon, R.T. Jr. (2000). The ecology and performance trade-offs in shell of freshwater molluscs. Cambridge armature of New Zealand mudsnails. University Press, UK. Ecology 87: 1038–1047. Dillon, R.T. Jr, Wethington, A.R., Holomuzki, J.R. & Biggs, B.J.F. (2007). Rhett, J.M. & Smith, T.P. (2002). Physical microhabitat effects on Populations of the European 3-dimensional spatial variability of freshwater pulmonate Physa acuta the hydrobiid snail, Potamopyrgus are not reproductively isolated from antipodarum. New Zealand Journal of American Physa heterostropha or Physa Marine and Freshwater Research 41: integra. Invertebrate Biology 121: 357–367. 226–234. Holomuzki, J.R., Lowe, R.L. & Biggs, Dole-Olivier, M-J., Marmonier, P. B.J.F. (2009). Spatiotemporal & Beffy, J-L. (1997). Response of separation of New Zealand mudsnails invertebrates to lotic disturbance: is from predatory fish. Journal of the the hyporheic zone a patch refugium? North American Benthological Society Freshwater Biology 37: 257–276. 28: 846–854. Dybdahl, M.F. & Kane, S.L. (2005). Jowett, I.G., Richardson, J., Biggs, B.J.F., Adaptations vs. phenotype plasticity in Hickey. C.W. & Quinn J.M. (1991). the success of a clonal invader. Ecology Microhabitat preferences of benthic Holomuzki & Biggs: Predator avoidance by invasive and native snails 23

invertebrates and the development of Journal of the North American generalized Deleatidium spp. habitat Benthological Society 7: 29–36. suitability curves, applied to four New McCarter, N.H. (1986). Food and energy Zealand streams. New Zealand Journal in the diet of brown and of Marine and Freshwater Research 25: from Lake Benmore, New Zealand. 187–199. New Zealand Journal of Marine and Kelly D.J. & Hawes I. (2005) Effects of Freshwater Research 20: 551−559. invasive macrophytes on littoral-zone McCarthy, T.M. & Fisher, W.A. productivity and foodweb dynamics (2000). Multiple predator-avoidance in a New Zealand high-country behaviours of the freshwater snail lake. Journal of the North American Physella heterostropha pomila: responses Benthological Society 24, 300–320. vary with risk. Freshwater Biology 44: Kristensen EA, Closs GP (2008). 387–397. Environmental variability and Peckarsky, B.L., McIntosh, A.R., Taylor, population dynamics of juvenile B.W. & Dahl J. (2002). Predator brown trout (Salmo trutta) in an chemical cues induce changes in upstream and downstream reach of a mayfly life history traits: a whole- small New Zealand river. New Zealand stream manipulation. Ecology 83: Journal of Marine and Freshwater 612−618. Research 42: 57−71. Pennuto, C. & Keppler, D. (2008). Short- Levine, J.M., Vilà, M., D’Antonio, C.M., term predator avoidance behaviour by Dukes, J.S., Grigulis, K. & Lavorel, invasive and native amphipods in S. (2003). Mechanisms underlying the Great Lakes. Aquatic Ecology 42: the impacts of exotic plant invasions. 629–641. Proceedings of the Royal Society of Rounick, J.S. & Winterbourn, M.J. London 270: 775–781. (1983). The formation structure and Levri, E.P. (1998). Perceived predation utilization of stone surface organic risk, parasitism, and the foraging layers in two New Zealand streams. behavior of a freshwater snail Freshwater Biology 13: 57–72. (Potamopyrgus antipodarum). Canadian Sagar, P.M. & Glova, G.J. (1995). Prey Journal of Zoology 76: 1878–1884. availability and diet of juvenile brown Levri, E.P., Kelly, A.A. & Love, E. (2007). trout (Salmo trutta) in relation to The invasive New Zealand mud snail riparian willows (Salix spp.) in three (Potamopyrgus antipodarum) in Lake New Zealand streams. New Zealand Erie. Journal of Great Lakes Research Journal of Marine and Freshwater 33: 1–6. Research 29: 527–537. Levri, E.P. & Lively, C.M. (1996). Schrieber, E.S.G., Glaister, A., Quinn, The effects of size, reproductive G.P. & Lake, P.S. (1998). Life history condition, and parasitism on the and population dynamics of the foraging behaviour in a freshwater exotic snail Potamopyrgus antipodarum snail, Potamopyrgus antipodarum. (Prosobranchia: Hydrobiidae) in Animal Behaviour 51: 891–901. Lake Purrumbete, Victoria, Australia. Lowe, R.L. & Hunter, R.D. (1988). Australian Journal of Marine and Effects of grazing by Physa integra Freshwater Research 49: 73–78. on periphyton community structure. Schreiber, E.S.G., Quinn, G.P. & 24 New Zealand Natural Sciences 37 (2012)

Lake, P.S. (2003). Distribution of C.M. (2000). Chemical cues modify an alien aquatic snail in relation to species interactions: the ecological flow variability, human activities and consequences of predator avoidance by water quality. Freshwater Biology 48: freshwater snails. Oikos 88: 148–158. 951–961. Turner, A.M., Fetterolf, S.A. & Bernot, Shimada, K. & Urabe, M. (2003). R.J. (1999). Predator identity and Comparative ecology of the alien consumer behavior: differential effects freshwater snail Potamopyrgus of fish and crayfish on the habitat use antipodarum and the indigenous snail of a freshwater snail. Oecologia 118: Semisulcospira spp. Venus 62: 39–53. 242–247. Sokal, R.R. & Rohlf, F.J. (1995) Biometry. Turner, A.M. & Montgomery, S.L. 3rd ed. New York, WH Freeman. (2003). Spatial and temporal scales of 468 p. predator avoidance: experiments with Statzner, B. (1981). The relationship fish and snails.Ecology 84: 616–622. between “hydraulic stress” and Wilhelm, F.M., Closs, G.P. & Burns, C.W. microdistribution of benthic (2007). Seasonal diet and amphipod macroinvertebrates in a lowland size selection of juvenile common bully, running water system, the Gobiomorphus cotidianus, in a coastal Schierenseebroks (North Germany). New Zealand lake. Hydrobiologia 586: Archiv für Hydrobiologie 91: 192–218. 303–312. Strayer, D.L. (2009). Twenty years Winterbourn, M.J. (1973). A guide to the of zebra mussels: lessons from the freshwater of New Zealand. mollusc that make headlines. Frontiers Tuatara 20: 141–159. in Ecology and the Environment 7: Winterbourn, M.J. & Fegley, A. (1989). 135–141. Effects of nutrient enrichment and Suren, A.M. (2005). Effects of deposited grazing on periphyton assemblages in sediment on patch selection by some spring-fed, South Island streams. two grazing stream invertebrates. New Zealand Natural Sciences 16: Hydrobiologia 549: 205–218. 57–65. Talbot, J.M. & Ward J.C. (1987). Zaranko, D.T., Farara, D.G. & Macroinvertebrates associated Thompsen, F.G. (1997). Another with aquatic macrophytes in Lake exotic mollusk in the Laurentian Alexandrina, New Zealand. New Great Lakes: the New Zealand native Zealand Journal of Marine and Potamopyrgus antipodarum (Gray Freshwater Research 21: 199–213. 1843) (, Hydrobiidae). Temperton, V.M., Hobbs, R.J., Nuttle, Canadian Journal of Fisheries and T. & Halle, S. (2004). Assembly rules Aquatic Sciences 54: 809–814. and restoration ecology: bridging the gap between theory and practice. Island Press, Covelo, , USA. Turner, A.M. (1996). Freshwater snails alter habitat use in response to predators. Animal Behaviour 51: 747–756. Turner, A.M., Bernot, R.J. & Boes,