Experimental set-up
The experiment was run from May to August 2012 in a screenedin wet laboratory at the Baruch Marine Field Laboratory in Georgetown, South Carolina, USA. Organisms used in the experiment were collected from the adjacent Oyster Landing intertidal oyster reef in North Inlet estuary (33°200N, 79°100W).

To examine the effects of individual crab activity level on the functional response, we measured both the activity level and mussel consumption rate of individual crabs. Activity level was measured prior to consumption rate trials. We manipulated the presence of chemical cues from toadfish during measurements of both activity level and consumption rate to test how the presence of predation threat directly affects the functional response, and how threat could mediate the effects of activity level on the functional response. Specifically, both the activity level and consumption rate of individual crabs were measured under one of two predation threat treatments: toadfish chemical cue absent (n = 240 crabs) or toadfish chemical cue present (n = 207 crabs). Activity level measured under predation threat is a measure of boldness as defined in the animal personality literature (Carter et al. 2013). The consumption rate of individual crabs was measured at a single prey density rather than a range of prey densities to minimize the duration crabs were held in the laboratory, which could modify individual behaviour through conditioning (Butler et al. 2006). We ran the experiment in a complete block design and the following methods pertain to a single block of 4 day duration.

On the first day, 16 crabs (20-30 mm carapace width, CW) were collected from the high intertidal portion of the Oyster Landing reef. Mud crabs become important predators of adult bivalves in oyster reefs in North Inlet estuary when they reach c. 20 mm CW (Toscano & Griffen 2012), and attain a maximum size of 55 mm CW at this site (McDonald 1982). All crabs were fed with mussels ad libitum as soon as they were brought into the laboratory. Eight of these 16 crabs were then randomly assigned to the toadfish cue absent treatment while the other eight were assigned to the toadfish cue present treatment, and these treatments were maintained for both activity level and consumption rate trials (methods for activity level and consumption rate trials are detailed below). To create the toadfish cue present treatment, we pumped seawater through a holding chamber that contained a single adult toadfish (c. 30 cm total length) fed ad libitum with mud crabs in between experimental trials. This seawater was then divided equally among mesocosms containing crabs to keep the amount of chemical cue consistent within blocks. Crabs assigned the cue absent treatment received seawater that did not first pass by a toadfish, but was otherwise distributed using the same seawater system. Mesocosms receiving the toadfish cue absent and cue present treatments were always alternated spatially.

On the second day, four crabs receiving the toadfish cue absent treatment and four crabs receiving the cue present treatment (eight of the 16 crabs) were observed to measure their activity level, and on the third day, the other eight crabs were observed in the same manner. This second group of eight crabs was fed again on the second day to keep their starvation time before activity level measurement (24 h) consistent with the first group of eight crabs, and on the third day, all crabs were fed to maintain starvation consistency before consumption rate trials. On the fourth day, the consumption rate of all 16 crabs was measured in a 24-h feeding trial. All crabs were held in the laboratory for an additional 2 days after consumption rate trials to ensure that crabs were not approaching a molt cycle or female crabs were not becoming reproductive. This procedure for a single experimental block was repeated 33 times over the course of the summer (May through August). Any crabs molting, carrying eggs or dying during their time in the laboratory were removed from the final data set. Additionally, the toadfish chemical cue treatment failed during the measurement of crab consumption rate for five blocks and these crabs were therefore removed from the final data set. However, complete removal of these blocks (i.e. both cue absent and present treatments) from the final data set did not alter our results.

Measurement of crab activity level
The activity level of individual crabs was measured using a similar behavioural assay to that used in previous studies of mud crab BTs (Griffen, Toscano & Gatto 2012; Toscano, Gatto & Griffen 2014). Each crab was observed in a glass mesocosm (50 x 28 x 30 cm) containing a 3 cm layer of sand/mud substrate and 5 L of oyster shell that had been dried and cleaned to remove epifauna. This experimental crab density (one crab per 0.14 m2) is within the range of densities previously reported in North Inlet (McDonald 1982, B.J. Toscano unpublished data). Oyster shell was placed on top of the substrate to mimic natural reef habitat. This amount of shell ensured that crabs had ample space to hide completely. In each tank, eight large mussels (c. 25 mm shell length, SL) were suspended in a mesh bag near the surface of the water to release chemical cues and induce crab foraging behaviour while remaining out of reach of crabs.

Crabs were observed at night (from c. 2000 to 2300 h) under dim red light and from behind a blind to minimize disturbance. Over a period of 3 h, we observed whether crabs were exposed and active (vs. hiding and remaining motionless) every 6 min (30 observations per crab in total). Activity level was measured as the proportion of 30 observations that crabs were visible to the observer and moving. In addition to activity level, we recorded the carapace width, major claw width and sex of each crab.

Measurement of crab consumption rate
Eight mussel (12-16 mm SL) prey densities (2, 4, 6, 8, 12, 16, 24 and 36 mussels per mesocosm) were randomly assigned to the eight crabs receiving the toadfish cue absent treatment as well as the eight crabs receiving the cue present treatment for each block. These mussel densities fall within the range of recorded mussel densities within a single large oyster cluster from the study site (Toscano & Griffen 2012). This created a total of 16 unique treatment combinations in each block. These trials were conducted in glass mesocosms of the same dimensions that we used to observe crab activity level. Mesocosms contained a 3 cm layer of sand/mud substrate and 10 large oyster shells, and were enclosed in black plastic to mimic the low-light conditions of North Inlet estuary during summer months (Dame et al. 1986; Toscano & Griffen 2013). Mussels were scattered evenly on oyster shells throughout each mesocosm and allowed to attach to oyster shells for 6 h prior to the start of trials. Crabs were allowed to forage for 24 h (starting and ending at c. 1500 h) and the number of mussels remaining as well as the water temperature was recorded at the end of trials.

Pherb_active_feeding.csv