When confronted with environmental stress, individuals may enhance their fitness by altering their phenotype via within-generation plasticity (WGP) or their offspring’s phenotype via transgenerational plasticity (TGP). Despite growing appreciation of TGP in adapting to environmental change, our understanding of how it may vary geographically among populations having different eco-evolutionary histories remains limited. In the Gulf of Maine, the carnivorous snail, Nucella lapillus, must solve the foraging-predation risk trade-off by reducing their vulnerability to invasive predatory green crabs while also acquiring the food necessary to survive. Nucella populations in the southern Gulf have been in contact with green crabs for ~125 years and are able to forage on an abundant supply of food. By contrast, populations in the northern Gulf have had limited contact with green crabs until ~25 years ago and must persist on a very limited food supply. In the southern Gulf, Nucella are known to exhibit TGP in response to green crab risk cues, but how these responses may vary along a spatiotemporal gradient in invasion history and food availability is unknown. We collected Nucella from six populations (3 northern, 3 southern) and evaluated WGP and TGP in response to green crab predation risk in an 18-month laboratory experiment. We detected regional variation in Nucella life history traits, the expression of both forms of plasticity in response to predation risk, and Nucella foraging on key basal resources. Our results support the hypothesis that low food availability in the northern Gulf may reduce the sensitivity of northern Nucella to predation risk. Hence, geographic variation in TGP in response to predation risk may be shaped by regional differences in other factors such as food availability, thereby yielding outcomes not well predicted by current TGP theory.

Methods

We conducted an 18-month laboratory experiment to examine geographic variation in predator-induced transgenerational plasticity of six Gulf of Maine Nucella populations (3 northern, 3 southern; see Appendix S1: Figure S1; Table S1 in associated manuscript for experimental design and site locations). We exposed parental snails to green crab predation risk (risk, no risk) and evaluated parental risk effects on offspring traits at emergence and the performance of 1-year old offspring under current risk (risk, no risk). To control for regional variation in water temperature (on average summer water temperatures in the northern Gulf are 6.62 °C cooler than in the southern Gulf; Corbett & Trussell, 2024), we conducted all parental exposure, breeding, offspring care, and the 1-year old offspring performance trial under ambient seawater temperatures typical of the southern Gulf in the flowing seawater system of the Northeastern Marine Science Center (MSC) in Nahant, MA.

Study Populations and Parental Risk Exposure

In February 2020, prior to the onset of Nucella mating season, we collected adult Nucella from six wave-exposed populations in the northern (N = 3) and southern (N = 3) Gulf of Maine (Table S1). Because the probability of sexual maturity at a given shell length differed among regions (Appendix S1: Figure S2; Table S2), we collected parents using region-specific size thresholds to avoid including sexually immature individuals. Indeed, at the 20 mm maturity threshold previously established in the southern Gulf (Donelan & Trussell 2018, a,b), we found that 93% of southern snails were mature while only 33% of northern snails were mature (Figure S2). Hence, we collected southern parents larger than 20 mm (23.17 mm ± 1.30; Mean ± SD) and northern parents larger than 25 mm (27.87 mm ± 1.30 ; Mean ± SD) to ensure all snails were sexually mature. After collection, we sexed Nucella and separated out 40 adult Nucella per population (N = 20 males, N = 20 females; 40 adults × 3 populations × 2 regions = 240 Nucella in total). We isolated each individual adult in a separate container (4.5 × 4.5 × 5.5 cm, L × W × H) and maintained them in seawater tables. Each adult Nucella was fed an ad libitum supply of blue mussels (Mytilus edulis, hereafter mussels; mean shell length = ~13 mm) for three months prior to the beginning of breeding trails. Separating each individual for three months prior to breeding was necessary to ensure Nucella paternity because females can store sperm from prior mating events for three months (Crothers, 1985). In May 2020, we randomly assigned adult Nucella to mating pairs from their respective populations (20 mating pairs per population) and then randomly assigned each breeding pair to a parental risk experience treatment (risk, no risk; Appendix S1: Figure S3). We exposed breeding pairs to risk in mesocosms (17 × 17 × 14.5 cm, L × W × H) that had one upstream compartment (12 × 12 × 10 cm, L × W × H) that held either a single adult male green crab and two Nucella for food (risk) or just had two Nucella (no risk). The mesocosms also had two separate downstream compartments (4.5 × 4.5 × 5.5 cm, L × W × H) each of which held either a single male or a female Nucella. We first fed parent Nucella mussels (N = 8 per individual) while they were exposed to their respective risk treatments (risk, no risk) for 4 days. Following this 4-day period, we then placed the male and female Nucella together in a separate risk-free container (8.5 × 8.5 × 10.5 cm, L × W × H) and allowed them to breed and lay egg capsules for 3 days. Allowing Nucella to lay eggs in the absence of risk cues regardless of parental risk treatment ensured that offspring within egg capsules did not experience embryonic induction. At the end of each 7-day breeding cycle, we isolated all egg capsules laid by a breeding pair in a mesh-lined tea infuser (5.5 × 5.5 × 6 cm, L × W × H). We then cycled parent Nucella between their risk treatment mesocosms and breeding containers for eight weeks (i.e., 8 cycles). We also recorded the number of mussels consumed by each parent and the clutch size laid by each breeding pair throughout the breeding trial. We analyzed mussel consumption and egg capsule production by parents with ANOVAs that considered predation risk exposure (risk, no risk) and region (north, south) as fixed effects and source population as a random effect nested within region. For mussel consumption our model also included sex (male, female) as a fixed effect and breeding pair as a random effect nested within region, source population, and predation risk exposure. We conducted all analyses in JMP 19 Pro (SAS Institute, 2025).

Offspring Traits at Emergence

To measure offspring traits at emergence, we monitored a subset of the egg capsules laid during the breeding trial. This subset of egg capsules included 4 families from each population x parental risk treatment combination (3 egg capsules × 4 families × 2 parent risk treatments × 6 populations = 144 total egg capsules). We maintained each individual egg capsule in its own tea infuser and checked for emerging juveniles every 3-4 days beginning six weeks after laying. We recorded the date of first emergence, counted the number of emerging juveniles, and photographed a subset of emerging juveniles (N = 378) with an AZ100 Nikon macroscope and NIS Elements Basic Research microscope imaging software (v. 4.30, Nikon Corporation, Tokyo, Japan). We analyzed offspring shell length and width at emergence with ANOVAs that considered parental risk exposure (risk, no risk) and region (north, south) as fixed effects, source population as a random effect nested within region, and family as a random effect nested within region, source population, and parental risk exposure. Because we measured the shell length and width of multiple individuals per egg capsule, we also included egg capsule as a random effect nested within region, source population, parental risk exposure, and family in our analyses.

We analyzed the number of days until emergence and the number of emerging juveniles with ANOVAs that considered parental risk exposure (risk, no risk) and region (north, south) as fixed effects, source population as a random effect nested within region, and family as a random effected nested within region, population, and parental risk exposure.

Culturing of Offspring

We maintained Nucella egg capsules from each parental pair in mesh lined tea infusers in the absence of risk as they developed. Six weeks after being laid, we began to monitor egg capsules every 3-4 days for hatching juveniles. Upon hatching, we fed offspring 200 juvenile mussels (~1 mm shell length). We then continued to feed Nucella a progressively larger ad libitum supply of mussels and transferred them to larger cylindrical (8.5 × 8.5 × 10.5 cm, L × W × H) containers until the offspring performance trial the following July (Donelan & Trussell, 2018a).

Offspring Performance Trial

In July 2021, we measured the tissue growth, shell growth, refuge use, and mussel consumption of 1-year old offspring in the absence and presence of current predation risk. We selected immature Nucella for use in our offspring performance trial based on population-specific size at sexual maturity data we collected. Importantly, due to the rapid growth of northern Nucella over the year-long culturing period, there were systemic differences among treatment groups in the size of immature individuals used in the trial. Northern Nucella began the offspring performance trial with approximately twice the tissue and shell mass than southern Nucella. While regional differences in offspring size may have modified the risk responses of Nucella, the patterns we observed among regions are inconsistent with past work (Matassa et al., 2016) that explicitly examined the influence of size on the foraging and growth of Nucella with and without predation risk (see Discussion). After selecting offspring and randomly assigning them to current risk exposure treatments, we estimated their initial shell and tissue mass using a non-destructive weighing technique (Matassa & Trussell, 2014; Palmer, 1982). We exposed offspring to current risk in mesocosms (27 × 15 × 5 cm, L × W × H) that consisted of two chambers separated by a perforated barrier. The upstream “risk” chamber (11 × 15 × 5 cm, L × W × H) was stocked with either one adult male green crab and 2 Nucella for food (risk) or just 2 Nucella (no risk). The downstream “response” chamber (16 × 15 × 5 cm, L × W × H) housed a granite tile (15 × 7.5 × 1 cm, L × W × H) that rested upon PVC spacers (1 cm H). This approach provided Nucella a choice between a refuge habitat (beneath the tile) that contained no food and risky habitat (the top of the tile) with food provided by 120 mussels (13.48 mm shell length (±) 0.02; mean (±) SE) for foraging Nucella to potentially consume. We stocked each response chamber with four Nucella offspring while being sure to include one individual from each family of the same parental experience x population combination. In total, we used 120 mesocosms for our offspring performance trial (5 replicate mesocosms × 2 current risk treatments × 2 parental risk treatments × 3 populations × 2 regions = 120 mesocosms; 4 snails per mesocosm × 120 mesocosms = 480 snails). For 28 days, we allowed offspring to grow and forage in mesocosms under their respective current risk treatment. We recorded the proportion of Nucella within each mesocosm in the refuge habitat every 3-4 days (Corbett & Trussell, 2024; Matassa et al., 2016) and collected mussels consumed by Nucella in each mesocosm every 7 days.

Upon termination of the offspring performance trial, we used a non-destructive weighing technique to estimate the tissue mass growth and shell mass growth of all Nucella offspring (Matassa & Trussell, 2014; Palmer, 1982). Due to differences in the initial shell and tissue mass of northern and southern Nucella, we conducted separate analyses for each region. We evaluated TGP in tissue growth and shell growth with ANCOVAs that considered parental risk treatment (risk, no risk) and current risk exposure (risk, no risk) as fixed effects, source population as a random effect, family as a random effected nested within source population and parental risk exposure, and mesocosm as a random effect nested within source population, parental risk exposure, and current risk exposure. Initial tissue or shell mass served as the covariate in these analyses.

We calculated refuge use as the proportion of snails sheltering in the refuge habitat (beneath the tile) within each mesocosm across all behavioral checks. The data for each region were analyzed with separate ANOVAs that considered parental risk treatment (risk, no risk) and current risk exposure (risk, no risk) as fixed effects, and source population as a random effect.

At the end of the offspring performance trial we measured every mussel consumed by Nucella to estimate total per capita energy acquisition using established methods (Matassa & Trussell, 2014). We focused on mean per capita consumption within each mesocosm because mussel consumption could not be accurately attributed to a single Nucella within a given mesocosm. For each region, we analyzed mussel consumption with separate ANOVAs that considered parental risk treatment (risk, no risk) and current risk exposure (risk, no risk) as fixed effects and source population as a random effect.