For manipulative experiments, we collected large brown crabs (10–12 cm carapace width, CW) by dredging the seafloor off Helgoland, small brown crabs (4–6 cm CW) using traps in the Helgoland rocky intertidal zone (54.177036, 7.884547) and lobsters (9–16 cm carapace length, CL; Supplementary Table S1) using traps off Helgoland. We purchased small (1 cm SL) and somewhat larger oysters (2 cm SL) at Morecambe Bay Oysters, medium-sized oysters (5–7 cm SL) at Bodes’s fish shop (Bremen, Germany) and large oysters (10–13 cm SL) at Rossmore Oysters (Stranraer, Scotland). We removed individual medium-sized and large Pacific oysters (Crassostrea gigas) and larger Pacific oyster clumps (9–15 cm SL) from low intertidal boulders in the Helgoland harbor (54.177178, 7.893714) using hammer and chisel and collected small mussels (Mytilus spp.) from pontons floating nearby (54.176808, 7.893872). We maintained all organisms in permanent seawater flow systems in the hatchery outdoor area. We fed crabs with large mussels (5–7 cm SL, five mussels/week) from subtidal ropes off Helgoland, lobsters with medium-sized Pacific oysters (five Pacific oysters/week) from the Helgoland harbor, and oysters with instant algae (Shellfish Diet 1800, Reed Mariculture Inc., Campbell, CA, United States; 250 mL/tank twice a week).

We conducted manipulative experiments in lab aquaria, lab mesocosms and outdoor tanks from June to November 2022 and from June to September 2023. Cuboid aquaria (31.5 cm × 18.5 cm × 16.5 cm, length × width × height, 9.6 L) were filled with 8 L water and received a constant flow of water (15 L/h). Mesocosms consisted of cylindrical basins (diameter: 126 cm, height: 50 cm, 620 L) containing 560 L water under constant flow (375 L/h) (Mackay-Roberts et al., 2024). Cuboid outdoor tanks (75 cm × 50 cm × 30 cm; 112 L) held 100 L water under constant flow (80 L/h). These dimensions ensured that all predators frequently encountered the offered prey. All aquaria and tank experiments were conducted under the natural light/dark rhythm. All mesocosm experiments were performed under red-filtered light, that does not influence brown crab and lobster activity (Conan et al., 1984; Davenport et al., 2023), with a 12 h light/dark rhythm. Basic information on all 19 experiments (hereafter: Exp_{1-17, S1-S2}) are summarized in Supplementary Figures S4, S5.

We visually confirmed that brown crabs and lobsters co-occur with oysters in the pilot oyster reefs, that lobsters interact with oysters in Helgoland harbor, that both predators interact with oyster spat-on-reef off Helgoland, that both predators crush and consume oysters in the lab, and that large lobsters consume medium-sized and large but not small oysters. We used one-sample t-tests to analyze one-choice experiments, two-choice experiments in which one oyster size (or prey) was not consumed and effects of conspecific presence on oyster consumption of crabs which did not consume oysters in conspecific presence. In all these cases, we compared the obtained mean number of consumed prey individuals (±SE) with zero. We used Student’s t-tests to analyze two-choice experiments and effects of lobster presence and waterborne predator cues on oyster consumption of brown crabs. Previously, we had checked whether the data met the assumptions for one-sample t-tests (normality) and Student’s t-tests (normality, variance homogeneity) using Kolmogorov-Smirnov and Levene tests (Sokal and Rohlf, 2012). In two cases, these assumptions were not met until the data were square root transformed after the constant “0.5” had been added (Sokal and Rohlf, 2012). When variance homogeneity could not be achieved through data transformation, we used Welch’s t-test for unequal variances (Ruxton, 2006). In one case, we omitted the one-sample t-test because the number of consumed oysters was very low (Figure 3D). In the Results section, we specified the type of each t-test when providing the corresponding result. We calculated Hedge’s g to compare the effect sizes of predator nonconsumptive effects and waterborne predator cues on oyster consumption of crabs. Hedge’s g ≥ 0.2 is a small effect, g ≥ 0.5 is a medium effect and g ≥ 0.8 is a large effect (Grizzard and Shaw, 2017). Since the duration of the experiments ranged between four and 264 h, we calculated standardized oyster consumption rates for 4 h to facilitate comparisons among the experiments (Figures 2, 3B–D, 5). We provide the raw data in Supplementary Tables S2–S23. We used Pearson correlation analyses to examine the shell thickness relationships between the left and right mussel valves (Section 2.3.2; Supplementary Figure S5) after we had confirmed that the shell thickness data met the required normality assumptions through Kolmogorov-Smirnov tests (Sokal and Rohlf, 2012). We performed these tests using GraphPad (www.graphpad.com) and Statistica (www.staftsoft.de) and plotted all data using Sigma Plot (www.grafiti.com).

All brown crabs and lobsters survived the experiments. Crabs and lobsters did not molt during the experiments. All crabs that were exposed to conspecific (Exp₉) or lobster presence (Exp₁₀) were intact and did not show any signs of predator attacks. In the controls, almost all oysters (99.9%) survived and oyster mortality was restricted to Exp₈ (Section 3.2.1; Figure 2F). Details on water temperature and salinity during the experiments are provided in Supplementary Figure S5.

The field surveys detected large brown crabs and a small lobster co-occurring in the two Borkum Reef Ground pilot oyster reefs (Figures 1A,B). Thus, mobile predators, that are considered early habitat colonizers (Kraufvelin et al., 2023; Smith et al., 2023) and typically occur on natural and artificial hard substrates (Krone et al., 2017; ter Hofstede et al., 2022; Krone and Schröder, 2011; Thatcher et al., 2023; Helgoland Lobster, 2022; Dybern, 1973), have colonized these reefs within 14 months (July 2020 to September 2021). Moreover, the occurrence of small brown crabs (4–5 cm CW) in oyster baskets (0.6 cm mesh width) next to the reefs (Section 3.1.1) indicates that crab megalopa larvae (0.2–0.4 cm CL (Ingle, 1981)) settling from the plankton (Sheehy and Prior, 2009) and the resulting tiny juvenile crabs (≥0.2 cm CW (Ingle, 1981)), that use such structurally rich habitats as protective nurseries (Mesquita et al., 2021), became recruited to the reef surroundings during that period. We also note that small and large lobsters were observed with the monitored oysters in Helgoland harbor (Section 3.1.2) and that large brown crabs and large lobsters were seen on (or near) both sandstone reefs and around oyster baskets kept on benthic landers off Helgoland 1 day and 22 days after sandstone reef deployment (Section 3.1.3). These findings show that early and later stages of oyster reef restoration projects are subject to predators from small to large sizes.

As lab experiments found that small brown crabs consume small oysters (Mascaró and Seed, 2001) and small to large brown crabs occurred in the pilot oyster reefs (Section 3.1.1), we examined whether these predators consume small to large oysters. Our experiments confirmed that small brown crabs consume small oysters (Figures 5C,D) and found, for the first time, that large brown crabs consume small (Figures 2A–C), somewhat larger (Figure 2B) and medium-sized (Figure 2D) but no large oysters (Figures 2C,D). Moreover, our field observations off Helgoland showed that very large brown crabs consumed even oyster spat-on-reef by chipping cemented oyster spat off the sandstone reef (Section 3.1.3). These findings show that brown crabs consume small to medium-sized oysters.

It was also examined whether predators interact with medium-sized oysters by deploying a camera-monitored oyster-lander on the Helgoland harbor seafloor (Section 2.1.2) which showed that at least one large lobster interacted with the oysters and consumed them in the field (Figure 1C). Interestingly, previous field observations in the Oslofjord (Norway) showed that lobsters transport their prey to their shelters for later consumption (Steen and Ski, 2014). However, as our camera was mounted relatively close to the lander, we could not determine whether all oysters were transported away from the lander by one or more lobsters and whether these oysters were consumed. Still, the three (i-iii) facts (i) that all six lobsters caught for the experiments in June 2023 were trapped within a relatively small area (approximately 25,000 m²) right outside Helgoland harbor within 1 day (Section 3.1.2), (ii) that we counted four large lobsters in proximity to each other near benthic landers off Helgoland in August 2023 (Section 3.1.3), and (iii) that territorial lobster individuals show overlapping home ranges elsewhere in the North Sea (Stamp et al., 2024; Moland et al., 2011; Skerritt et al., 2015) suggests that the Helgoland harbor lander was visited by more than one lobster. Similar to the aforementioned Oslofjord observations (Steen and Ski, 2014), we saw that the lobsters in the mesocosms collected oysters from the floors and stored them along the walls for later consumption (Section 3.2.1). We also observed one small lobster₁₃ feeding on small oysters in an outdoor tank (Section 3.2.1) and another small lobster attempting to feed on the oyster spat-on-reef off Helgoland (Section 3.1.3). Additionally, we note that most lobster-oyster interactions (77%) in the Helgoland harbor were observed under dim daylight conditions within 2 h after sunrise (7–9 a.m.) and sunset (7–9 p.m.; Supplementary Figure S12) (Sunrise and Sunset Helgoland, 2024) which corresponds with the current opinion that lobsters are no nocturnal but primarily crepuscular predators (Davenport et al., 2023). Overall, these findings show that lobsters interact with small to medium-sized oysters and consume them in the lab and in the field. Accordingly, predation by brown crabs and lobsters poses a risk for small to medium-sized oysters in restoration projects almost instantly after reef deployment.

Our lab experiments showed that large crabs consumed on average (±SE) 3.80 ± 1.06 (Figure 2C) to 8.30 ± 0.58 small oysters (Figure 2A) and 6.89 ± 1.02 somewhat larger oysters (Figure 2B) within 4 h. In contrast, large crabs consumed only 0.01 ± 0.01 oysters (2–5 cm SL; Figure 3D) and 0.01 ± 0.32 medium-sized oysters within the same time (Figure 2D). Thus, large crabs consumed 380 to 830 times more small oysters and 689 times more somewhat larger oysters than medium-sized oysters. We also repeatedly found that large crabs did not consume large oysters (Figures 2C, D), even over an 11-day experiment (Figure 2D; Supplementary Figure S4). These results indicate strong crab predation pressure on smaller oysters. Previous lab work with small brown crabs (2–4 cm CW) and small European oysters (0.3–3 cm SL) had shown that small crabs consumed small oysters (0.3–2.1 cm SL) but did not consume larger oysters (2.1–3 cm SL) (Mascaró and Seed, 2001). When such crabs were offered mussels, Pacific oysters or cockles (Cerastoderma edule) of similar sizes (0.2–3 cm SL), crabs consistently consumed smaller prey individuals (mussels: 0.4–2.4 cm SL; Pacific oysters: 0.3–2.4 cm SL; cockles: 0.2–2.1 cm SL) (Mascaró and Seed, 2001). Likewise, crabs (7.8–15 cm CW) offered horse mussels (Modiolus modiolus, 0.7–9.9 cm SL) preferred smaller horse mussels (1.5–5.9 cm SL) (Seed et al., 1975). Also, crabs (6.1–14.1 cm CW) fed scallops (Pecten maximus, 3.0–9.0 cm SL) preferentially consumed smaller scallops (3.0–4.0 cm SL) (Lake et al., 1987). Furthermore, crabs (7.6–8.6 cm CW) offered dogwhelks (Nucella lapillus, 1.0–2.5 cm SL) preferred smaller dogwhelks (1.0–1.5 cm SL), and crabs (6.4–9.8 cm CW) offered periwinkles (Littorina littorea, 1.0–3.0 cm SL) preferentially ate smaller periwinkles (1.0–2.0 cm SL) (Lawton and Hughes, 1985). All these preferences support the notion that brown crabs consistently select smaller bivalve and other molluscan prey to limit prey handling time to reduce risk of competition and predation (Juanes, 1992; Seed and Hughes, 1995) and to avoid claw wear and damage resulting from handling larger prey that can lead to energy loss, poorer agility and defense ability (Juanes, 1992; Juanes and Hartwick, 1990). Altogether, these results clearly show that brown crabs strongly prefer smaller oysters (and other small molluscs), whereas their consumption of medium-sized and large oysters is negligible.

Concerning lobsters, a positive (but flattening) correlation between predator size and prey size preference appears plausible. For example, we observed that a small lobster₁₃ readily consumed all offered small and somewhat larger oysters within the first night of both experiments, whereas large lobsters_14-18 did not consume such small oyster sizes over two consecutive 7-day experiments (Section 3.2.1; Figure 2E). We also found that large lobsters_1-12 strongly preferred medium-sized oysters over large oysters (Figure 2F). Similarly, small American lobsters (Homarus americanus, 5.5–9.8 cm CL) offered American oysters (1.0–3.5 cm SL) preferred small (1.0–2.5 cm SL) over larger oysters (Elner and Lavoie, 1983). Information on European lobster preferences for sizes in other prey species (e.g., bivalves and snails (Hallbäck and Warén, 1972)) seemingly does not exist. However, prey size preferences in other lobster species, including American, Californian (Panulirus interruptus), Caribbean (Panulirus argus), New Zealand (Jasus edwarsii) and South African lobsters (Jasus lalandii), often increase with lobster size (Elner and Jamieson, 1979; Griffiths and Seiderer, 1980; Tegner and Levin, 1983; Robles et al., 1990; James and Tong, 1998; van Zyl et al., 1998; Sainte-Marie and Chabot, 2002; Hanson, 2009; Eurich et al., 2014; Gnanalingam and Butler IV, 2018) which resembles our findings and likely reflects the changing ability of lobsters to handle prey up to a certain prey size as lobsters grow (Boudreau and Worm, 2012). Overall, these findings indicate that medium-sized to large oysters have reached a prey size refuge from brown crab but not lobster predation, and that large oysters experience lower lobster predation.

Large crabs offered mussels and slightly smaller oysters simultaneously (Figure 3A) consumed mussels but did not consume oysters (Figure 3B). Likewise, small crabs preferred small mussels over similar-sized oysters (Mascaró and Seed, 2001). These preferences for mussels are probably driven by mussels being a more profitable prey than oysters. For instance, our mussel shells were thinner than our oyster shells at the shell lip (Figure 3E), center (Figure 3F) and base (Figure 3G) and, thus, easier to crack. Mussels also have a somewhat higher energy content (range: 5.30–5.66 cal/mg ash-free dry weight; AFDW) than oysters (5.21 cal/mg AFDW) (Beukema et al., 1997) and our mussels contained 11 times as much flesh as our oysters (Figure 3H) indicating that our mussels were more nutritious than our oysters. When offered mussels and oysters separately, crabs consumed almost 15 times as many mussels as oysters within the same time period (Figures 3C,D). Furthermore, crabs started consuming mussels already on the first day of the experiment but did not consume any oysters before the third day (Section 3.2.2) showing that prey handling time for mussels is shorter than for oysters. Accordingly, we estimated that our mussels (0.8033 g/d; average tissue dry weight/estimated handling time of 1 day) were 34 times as profitable as our oysters (0.023311 g/d; average tissue dry weight/estimated handling time of 3 days). Similarly, preferences of invasive green crabs and native red rock crabs (Cancer irroratus) in Prince Edward Island (PEI, Atlantic Canada) for soft-shell clams (Mya arenaria) and mussels over American oysters (Miron et al., 2005; Pickering and Quijón, 2011) were negatively related to shell thickness differences among these three bivalve species (Pickering and Quijón, 2011). Finally, we note that small crabs did neither prefer small mussels over similar-sized cockles nor small Pacific oysters over similar-sized European oysters (Mascaró and Seed, 2001) and that lobsters (based on stomach content analyses) tend to prefer crab, snail, and polychaete prey over bivalve prey (Hallbäck and Warén, 1972) but did not prefer medium-sized Pacific oysters over similar-sized European oysters (Figure 4D). Considered together, these findings indicate that prey profitability drives brown crab preferences for prey and that the presence of alternative prey, such as mussels, cockles and Pacific oysters, can reduce predation pressure on European oysters.

Individual oysters can form oyster clumps by overgrowing conspecifics (Reise et al., 2017a; Merk et al., 2020). Such oyster clumps are typically larger than individual oysters and may, thus, have reached a prey size refuge from predation. To test this prediction, we conducted an experiment in which lobsters were offered individual medium-sized Pacific oysters and larger Pacific oyster clumps (Section 2.3.3). We used Pacific oysters, which are widespread in the North Sea (Troost, 2010; Reise et al., 2017b) and common around Helgoland (Zwerschke et al., 2013; Ellrich et al., 2023), as model organisms because European oyster clumps were not available. Our 6.5-day experiment found that lobsters preferred individual oysters over oyster clumps (Figure 4A) and that lobsters barely consumed clumped oysters. Actually, only a few oyster clumps showed traces of lobster attacks during which only oysters protruding from the clumps were consumed (Figure 4A) likely as the clumps were larger and heavier than the individual oysters (Figures 4B, C) and, thus, more difficult for the lobsters to handle. These results indicate that oyster clumps are relatively safe from large predators. Interestingly, paired juvenile American oysters (i.e., two small individuals grown together) in PEI experienced lower predation by invasive green crabs than individual juvenile American oysters (Poirier and Quijón, 2022). We also found that lobsters do not prefer medium-sized Pacific over similar-sized European oysters (Figure 4D) suggesting that European oyster clumps are also safer from predation than individual European oysters which should be examined through future manipulative experiments.

As cannibalism occurs among brown crabs (Lawton, 1989; Amaral et al., 2009), lobsters prey on brown crabs (Hallbäck and Warén, 1972) and predator presence often reduces prey feeding activity since prey typically ceases feeding to decrease predation risk (Trussell et al., 2003; Molis et al., 2011; Pessarrodona et al., 2019; Catalán et al., 2021; McCall et al., 2023; Curtis and Wing, 2024), we examined whether such nonconsumptive effects (NCEs) by conspecifics and lobsters would reduce oyster consumption of crabs (Section 2.3.4). Working with a crab abundance in the mesocosms (2.4 crabs/m²) which closely corresponded with the natural crab abundance in the pilot oyster reefs (2.0–3.1 crabs/m²; Section 3.1.1), we found that conspecific presence essentially “switched off” crab consumption of medium-sized oyster (Figure 5A). This shows that the presence of cannibalistic conspecifics heavily diminishes oyster consumption of crabs likely as the crabs spend more time avoiding conspecifics when detecting conspecific cues. Interestingly, mutualistic interferences among interacting conspecifics can also limit prey consumption of brown crabs and, thereby, cause a disproportional increase in prey consumption with crab density (Amaral et al., 2009). Thus, predator NCEs (this study) and mutualistic interferences (Amaral et al., 2009) are two different mechanisms underlying such a disproportional increase that ultimately facilitates prey survival. In contrast, such a disproportional increase may not occur among non-cannibalistic predators. This notion is supported by field experiments which showed a positive relationship between the number of adult predatory dogwhelks (that only show slight cannibalistic tendencies as juveniles (Largen, 1967; Crothers, 1985; Gul’bin and Shadrin, 1991; Gosselin and Chia, 1994)) and the number of mussels consumed by them (Meister et al., 2023). The presence of only one lobster also limited brown crab consumption of medium-sized oysters by 77% (Figure 5B) which indicates that even individual lobsters, like the one in the eastern pilot oyster reef (Figure 1B), contribute to nonconsumptive predator regulation and limitation of brown crab feeding activity in the field. Finally, we detected that conspecific presence (Hedge’s g = 1.137) and lobster presence (g = 1.334) had similarly large negative effects on oyster consumption of crabs. However, as predator NCEs on prey typically increase with predator abundance (Silberbush and Blaustein, 2011; Hill and Weissburg, 2013; Ellrich et al., 2015; Büchner-Miranda et al., 2024) and both mesocosm experiments (Exp_{14, 15}) used different predator abundances (i.e., three crabs/mesocosm versus one lobster/mesocosm), these findings suggest that individual crab NCEs are likely somewhat weaker than the detected individual lobster NCEs. Overall, these results show that brown crab and lobster presence can regulate and limit oyster consumption of brown crabs under natural predator abundances and suggest (in combination with the strong brown crab preferences for small and somewhat larger oysters; Section 4.3) that medium-sized and larger oysters are safe from brown crab predation when exposed to such natural brown crab and lobster abundances.

Waterborne cues released by brown crabs and lobsters trigger defense responses in mussels (Côté, 1995; Côté and Jelnikar, 1999; Wang et al., 2024). Therefore, we examined whether the detected conspecific and lobster NCEs on oyster consumption of brown crabs (Figures 5A, B) are mediated by such cues. For that, we measured oyster consumption of brown crabs in conspecific water, lobster water and water (Section 2.3.5). Conspecific and lobster water reduced oyster consumption of crabs by 81% and 89%, respectively (Figures 5C, D), indicating that waterborne conspecific and lobster cues mediate the detected nonconsumptive regulation and limitation of crab predation. Conspecific and lobster water had large effects on oyster consumption of crabs (conspecific water: Hedge’s g = 1.013; lobster water: g = 1.406). Yet, predator NCEs on prey intensify with predator cue concentration (Büchner-Miranda et al., 2024; Loose and Dawidowicz, 1994; von Elert and Pohnert, 2000; Ferland-Raymond et al., 2010) and we used three crabs but only one lobster to produce conspecific and lobster water (Section 2.3.5). This suggests, as the aforementioned effects of conspecific and lobster presence on crab predation (Section 4.6), that individual crab NCEs are likely somewhat weaker than the detected individual lobster NCEs. This conclusion is corroborated by our two pilot studies which found that crab water, that derived from three crabs, had a weaker effect on mussel byssal thread production (g = 1.373, Supplementary Figure S1A) than lobster water from an individual lobster (g = 1.884, Supplementary Figure S1B). Finally, predator NCEs on prey can also be triggered by tactile and visual predator cues (Tapia-Lewin and Pardo, 2014; Boudreau et al., 2018). However, the effects of conspecific presence (g = 1.137) and conspecific water (g = 1.013) as well as lobster presence (g = 1.334) and lobster water (g = 1.406) on oyster consumption of crabs were quite similar which further corroborates that the detected predator NCEs were mediated by waterborne predator cues.

Although our study focused on consumptive and nonconsumptive interactions among common predators and European oysters in the North Sea, we note that brown crabs and lobsters are NE Atlantic predators that occur from northern Norway (Europe) to Western Sahara (northern Africa) (González, 2022; FAO, 2023a) which encompasses the entire historic European oyster distribution range from southern Norway to southern Spain (Pogoda, 2019; Thurstan et al., 2024) where several oyster reef restoration projects are currently underway (Pogoda, 2019; Pogoda et al., 2024; zu Ermgassen et al., 2021). Accordingly, our findings provide fundamental knowledge of predator-prey interactions that are relevant for oyster reef restoration in Europe.

Prey size refuges may protect oysters also from other common shell crushing NE Atlantic predators such as juvenile green crabs, that interacted with juvenile oysters and preferred small over large juvenile oysters in lab experiments (Mascaró and Seed, 2001), and swimming crabs (Liocarcinus spp.), that prey on thin-shelled bivalves (Choy, 1986) and were, for the first time, observed attempting to prey on oyster spat-on-reef off Helgoland (Section 3.1.3). This conclusion is corroborated by the seven (i–vii) findings in North America which indicated (i, ii) that larger American oysters (≥5.5 cm SL) showed lower mortalities when exposed to invasive green crabs and native red rock crabs than smaller American oysters (≤5.5 cm SL) (Poirier et al., 2017), (iii, iv) that large American oysters (2.5–4.0 cm SL) were not consumed by these two predators when offered together with small (≤1.5 cm SL) and medium-sized American oysters (1.5–2.5 cm SL) (Miron et al., 2005), (v) that medium-sized (2.5–3.5 cm SL) and large American oysters (3.5–5.5 cm SL) showed lower mortalities under green crab predation than small American oysters (1.5–2.5 cm SL) (Pickering et al., 2017; Tummon Flynn et al., 2015) with (vi) large oysters not being consumed by green crabs (Pickering et al., 2017), and (vii) that medium-sized American oysters (2.5–3.5 cm SL) suffered lower mortality from green crab predation than small American oysters (1.5–2.5 cm SL) (Poirier and Quijón, 2022). Accordingly, larger European oysters are likely safe from predation by green crabs and swimming crabs that have similar claw morphologies, claw strengths and shell crushing methods (ap Rheinallt and Hughes, 1985). Also, the presence of brown crabs and lobsters, that prey on green crabs and other predatory crab species (Hallbäck and Warén, 1972; Lawton, 1989) and show interspecific agnostic behaviour towards other predatory crab species (Davenport et al., 2023), may regulate and limit European oyster consumption by other predatory crab species, such as green crabs, through predator consumptive and nonconsumptive effects.

Our results also provide important information for other shellfish restoration projects in Europe (Carranza and zu Ermgassen, 2020; Fitzsimons et al., 2019) focusing, for example, on declining horse mussel populations (Strong et al., 2015; Fariñas-Franco and Roberts, 2017) in Denmark (BioReef, 2024) and Northern Ireland (Modiolus Restoration Research Group, 2024) whose subtidal mussel beds are biodiversity hotspots across northern European waters (Fariñas-Franco et al., 2023). Brown crabs and lobsters consume horse mussels (Hallbäck and Warén, 1972; Shelton et al., 1979; Seed et al., 1975) and brown crabs prefer small over larger horse mussels (Seed et al., 1975). Thus, prey size refuges and predator consumptive and nonconsumptive effects may limit predation on other shellfish habitats under restoration, such as horse mussel beds, as well.

Finally, as green crabs and red rock crabs have cannibalistic tendencies (Ojeda and Dearborn, 1991; Stehlik, 1993; Baeta et al., 2006; Chaves et al., 2010; Rondeau et al., 2014; Gehrels et al., 2017; Geburzi et al., 2018) and green crab ethograms show intraspecific agonistic behaviour and intraspecific kleptoparasitism (Davenport et al., 2023), it would be interesting to investigate whether mutualistic interferences among interacting conspecifics (Amaral et al., 2009), conspecific presence and waterborne cues (this study) also regulate and limit American oyster consumption by green crabs and red rock crabs and, thereby, facilitate American oyster survival in the field.

Based on this study, we recommend to identify predator-prey relationships and the corresponding predator-prey interactions to understand offshore reef community ecology and to develop and test measures regulating predation in ecological reef restoration. As first steps, literature reviews and field surveys should determine predators, prey, alternative prey and top-predators (that prey on predators) occurring in the respective region. In Europe, these can include, for example, green crabs and brown crabs (predators), oysters (prey), mussels (alternative prey) and lobsters (Mascaró and Seed, 2001, this study) or large fish such as skates or rays (top-predators) (Poiesz et al., 2021). Next, field observations and manipulative experiments should establish the predator hierarchy and corresponding predator-prey interactions (this study). Building on this, measures for natural regulation of predation (increasing in complexity concerning feasibility and efficiency) can be developed: from one-species (oyster) over two-species (oyster, mussel) to multi-species approaches (oyster, mussel, lobster, fish). To test and compare these approaches, we suggest to monitor them with regard to reef development over time (Pineda-Metz et al., 2023; zu Ermgassen et al., 2021).

For the one-species approach and as oyster reef building blocks, we recommend to use larger individual oysters which have reached a size refuge protecting them from predators (i.e., ≥7 cm oyster SL for brown crabs, ≥12 cm oyster SL for lobsters, Section 4.3) and oyster spat-on-large-shells or oyster clumps (spat-on-clumps) as oyster size (Section 4.3) and clumping reduced predation pressure (Section 4.5). These larger oyster sizes certainly mean higher economic costs due to more effort and longer production phases for oyster hatcheries and nurseries. However, predators consistently preferred crushing the shells of small oysters and other molluscan prey (Sections 4.3, 4.5, 4.8). Accordingly, larger single oysters, spat-on-large-shells and spat-on-clumps are likely better protected from shell-crushing predators than small and single seed oysters. Using them in restoration projects, should, therefore, increase oyster survival in the field and oyster reef restoration success in the long-term.

Concerning the two-species approach, we suggest to add an alternative prey, such as mussels (Section 4.4), if native to the restoration region. As mussel beds as biodiversity hotspots are also declining (Meister et al., 2023; Banke et al., 2024; Little et al., 2024; Cameron and Scrosati, 2023), the ecological restoration of both systems, oyster reefs and mussel beds, can go hand in hand where applicable and ecologically relevant. Some predators should prefer mussels as more profitable alternative prey over oysters (Section 4.4), so that the alternative prey can distract and nourish these predators and, thereby, regulate and limit predation pressure on oyster reefs.

Finally, regarding the multi-species approach, we recommend to include native top-predators (if not already present), such as European lobsters or large fish, such as skates, that prey on predators, tend to prefer crab, snail and polychaete prey over bivalve prey (Section 4.4) and limit oyster consumption of predators, in the case of lobsters also through predator NCEs (Section 4.6) mediated by waterborne predator cues (Section 4.7).

We conclude that common brown crabs and European lobsters are predators of European oysters and that oyster size refuges, the presence of alternative prey, oyster clumping, as well as intra- and interspecific regulation and limitation of oyster consumption by predators mediated through predator NCEs, that are triggered by waterborne predator cues, can regulate and reduce predation pressure on European oysters. Since brown crabs and lobsters co-occur across the entire historic European oyster distribution range (from southern Norway to southern Portugal) where several oyster reef restoration projects are conducted, our study provides fundamental knowledge of predator-prey interactions that is required for developing measures to regulate and reduce oyster predation and increase oyster reef restoration success in European waters.