Dataset: P. herbstii - refuge use (BESB, 2014)

We first surveyed the individual refuge use of 247 mud crabs in the absence of a toadfish chemical cue and the refuge use of 224 separate crabs in the presence of the cue (i.e., under predation threat) from May to August 2012. All crabs were collected by hand from an oyster reef known as Oyster Landing in North Inlet estuary, Georgetown, SC, USA (33°20'N, 79°10'W). Crabs were collected within a 20 m× 20 m area at the center of the reef. Behavioral measurements were made in a screened-in wet laboratory at the adjacent Belle W. Baruch Institute for Marine and Coastal Sciences. Measuring crab refuge use behavior in the field was not possible due to the high turbidity of water in North Inlet during the summer months that limits visibility (Dame et al. 1986).

Initial behavioral measurements:
The following describes our procedure for a single observational block. Thirty-two observational blocks were run over the course of the study (May-August). Sixteen crabs were collected between 20 and 30 mm carapace width (CW) from Oyster Landing reef, and we attempted to ensure that each collection reflected the entire crab size range (20-30 mm CW). We randomly assigned eight of these 16 crabs to the toadfish cue absent treatment and the other eight to the toadfish cue present treatment. Due to constraints on the number of crabs a single person could observe in a night, we observed eight crabs per night over two consecutive nights, generally from 2000 to 2300 h. During each night, four crabs receiving the no cue treatment and four crabs receiving the cue treatment were observed. The night measured (first or second) had no effect on refuge use behavior (ANOVA: p >0.05), so the blocking factor used in our analyses was the 2-day span over which 16 crabs were measured. Any crabs molting, carrying eggs or dying during their time in the lab were removed from the data set.

Refuge use was measured following the behavioral assay protocol used in Griffen et al. (2012). All crabs were starved for 24 h before their refuge use behavior was measured. Each crab was observed in a separate glass mesocosm (50 cm× 28 cm×30 cm) containing a 3 cm layer of sand/mud substrate and 5 L of oyster shell (8-12 cm shell length) that had been dried and cleaned to remove epifauna. This amount of shell ensured that crabs had ample refuge to hide completely. Mesocosms were completely filled with a continuous supply of seawater. Eight large scorched mussels were suspended near the water surface in a mesh bag to release prey chemical cues and induce crab searching behavior while remaining out of reach of crabs. To create the toadfish cue treatment, crabs received a continuous supply of seawater that was first pumped through a holding chamber that contained a single adult oyster toadfish. Crabs assigned the no cue treatment received a continuous supply of seawater pumped through a holding chamber without a toadfish.

Crabs were observed under red light with the observer located behind a blind to minimize crab disturbance. Crabs were first given a 15-min acclimation period in the observation tanks, after which their refuge use was observed once every 6 min over 3 h (30 observations in total for each crab). Refuge use was measured as the proportion of the 30 observations where crabs were completely in the oyster shell refuge and thus invisible to the observer. The refuge was a matrix of shells, so crabs could be hiding under a single shell or multiple shells (i.e., at the bottom of the matrix). When crabs were observed out of the refuge, they were usually walking on top of the shell substrate (i.e., were active). In addition to refuge use behavior, we measured the carapace width and sex of each crab.

Repeatability and behavioral change:
After refuge use behavior was measured, each crab was marked with a unique ID number and released back into the field. To mark crabs, we glued (with super glue) a piece of laboratory labeling tape numbered with permanent marker to the center of the crab's carapace (Stachowicz and Hay 1999). We realized during the middle of the study that these handwritten numbers were becoming illegible over time in the field, and so the last 80 crabs from the survey were marked using plastic bee tags (queen marking kit: the Bee Works, Orillia, Ontario, Canada) that did not wear over time. All marked crabs were released in a 10 m×10 m area at the center of Oyster Landing reef. To assess the persistence of refuge use behavior for crabs released into the field, we recaptured the crabs and measured their refuge use for a second time using the same behavioral assay procedure in the lab. Recaptured crabs were observed under the same treatment (toadfish cue absent or present) that they were observed under before release. We recaptured the crabs by hand within the release area of Oyster Landing reef. We searched the reef over two separate search periods (end of July and end of August) until no more marked crabs were recovered at each search period. Because crabs were released regularly over the duration of the study (after each block) but resampled just twice, individual crabs were recaptured after different durations in the field, allowing us to test the effects of duration in the field on behavioral change over time.