Dataset: P. herbstii mortality data

We performed an experiment to assess the proportion of time individual mud crabs spend active vs hiding in refuge in the presence of predator odor cues from either a single toadfish, a single blue crab, or control conditions with no predator cue. Previous work by Toscano et al. (2014) determined that differences in this behavior between individuals persist over months (i.e. the behavior defines the personality of individuals). We collected 300 mature mud crabs that were not missing any limbs (mean ± SD carapace width = 24.1 ± 2.2 mm) by hand from intertidal oyster reefs within the North Inlet National Estuarine Research Reserve (33°20’N, 79°10’W, Georgetown, South Carolina). Crabs were collected in three cohorts of ten individuals during each of 10 blocked sampling periods (i.e., each blocked trial consisted of 30 crabs total). Individuals were randomly selected from 1 m² plots to ensure that natural cohorts of ten crabs were measured. Crab gender was identified by examination of the telson (167 males, 133 females).

Crabs were starved for 24 h and the carapace of each was marked with a unique nail polish (Sonia Kashuk) design to identify individuals. Preliminary work determined that these markings did not alter crab behavior. Cohorts were randomly assigned a predator cue treatment and placed into one of three separate flow-through mesocosms (circular with diameter 1 m; water height 15 cm). Each mesocosm contained ~2 cm sediment under a ~8 cm matrix of cleaned oyster shells covering the entire tank bottom. Thirty scorched mussels were distributed in three mesh containers within each mesocosm outside the reach of crabs to continuously stimulate foraging behavior [20]. Flow-through mesocosms were supplied with water from the estuary which was first pumped through a head tank. The head tank for each mesocosm contained either a mature toadfish (caudal length ± SD = 28.4 ± 2.8 cm), blue crab (carapace width ± SD = 14.8 ± 0.6 cm), or no predator depending on the predator odor cue treatment. Predators were caught from the estuary by dip net no more than one week prior to the experiment and fed mud crabs each day to ensure kairomones were produced. We conducted all experimental trials at night under red light following the observational procedures of Griffen et al. (2012) and Toscano et al. (2014)  to ensure mud crabs were at their most active and were undisturbed by the observer. 

Crabs were tested in trials consisting of one cohort per treatment (toadfish, blue crab, no predator) with six days separating the commencement of each trial. All trials began between the hours 2000-2100, and once cohorts were placed in the mesocosms, crabs were given 10 minutes to acclimate. After acclimating, we recorded whether crabs were actively exposed on the surface of the shell layer or were taking refuge underneath the shells at six minute intervals for the next three hours. The proportion of these 30 observations in which crabs hid in refuge and were not visible to the observer was used as our response variable.

Immediately after observing crab behavior (described in the last subsection), we used the same crab cohorts to assess whether crab predation risk was influenced by the proportion of time individual crabs spend in refuge within oyster shells and by predator species. We assigned crabs the same predator treatment they experienced previously to keep cohorts intact throughout the entire study.  Crabs were fed a satiating amount of fish (Fundulus heteroclitus), marked with individually-numbered bee tags (queen marking kit: the Bee Works, Orillia, Ontario, Canada), and starved for 24 h. After the starvation period, the cohorts were placed into one of three large flow-through mesocosms (diameter 2m; water height 90 cm). Each mesocosm contained ~ 2 cm of sediment underneath four clusters of live oysters (length ~38 cm, width ~31 cm, height ~28 cm) which were standardized by weight (15.000 kg within <0.1 %). Oysters were collected from the estuary and cleaned of any inhabiting crabs. Scorched mussels naturally attached to these collected oysters were standardized by number of individuals (within 8.3 %) and served as the mud crabs’ food source to mimic natural conditions.

During each blocked trial (n=10), a single toadfish, blue crab, or no predator (to serve as a control for cannibalism) was placed in each mesocosm depending on the experimental treatment. We used the same individual predators which provided odor cues in the previous behavior experiment as predators within the large mesocosms. Predators were starved 24 h to standardize hunger levels and placed in the mesocosms 10 minutes prior to the mud crabs to ensure kairomones were distributed throughout the tank (no crabs were lost to predation during introduction into the tanks). Mud crab survival was checked daily for seven consecutive days to determine which individuals were consumed.  This was done by removing all the oyster clumps and thoroughly raking the sediment. Any missing crabs were presumed dead as there was no way for crabs to escape and remnants of missing animals were often found.