In the summer of 2015 we quantified predation risk for mesograzers using identical experiments conducted at 17 sites spread across North America, Europe, and Asia (see Table 1 at end of this document). In each experiment, we used tethering to determine relative predation risk for locally collected organisms along patch edges and in patch interiors under three levels of simulated eelgrass degradation (0, 50, and 80% shoot loss) in a crossed design. Epifauna selected for tethering were small (approximately 2 – 20 mm in length) mesograzers (gammarid amphipods and isopods) that are commonly found in the guts of small fishes (Table 1). Tethering measures the relative mortality rate of prey among different treatments, and because tethered prey cannot flee from predators, represents the relative mortality rate for prey that are readily available to predators (Aronson & Heck 1995). Though organisms used in tethering experiments differed in species and sizes among sites, we only used taxa of sizes that are commonly found in the guts of predators at each site.

To set up experiments, at each site we first selected a large eelgrass bed (typically > 5,000 m2) in shallow water (0.5 – 1.5 m water depth at low tide) with a distinct edge formed by an abrupt transition from eelgrass to unvegetated sand or mud. Edge habitat was defined as being within eelgrass but within 1 m of the transition from eelgrass to unvegetated sediment, and interior habitat was > 5 m from this transition. We chose these distances because in seagrass habitat edge effects on mortality and abundance of small epifauna typically occur within 1 m from patch edges (Tanner 2005, MacReadie et al. 2010). Patch vegetation consisted exclusively of eelgrass, except for epibionts or sparse drift algae. At each site, we created 21 experimental blocks along the edge and 21 experimental blocks within the interior of the eelgrass bed. Each block consisted of two 1 m x 1 m eelgrass plots separated by a distance of 30 cm. One randomly selected plot in each block was designated for tethering mesograzers, and the other plot was used to tether larger organisms in a companion experiment (data are not listed for this companion experiment). We randomly selected seven of the 21 blocks at the edge and in the interior, and after obtaining shoot counts within these plots, haphazardly pulled shoots by hand to thin each plot to 50% of its ambient shoot density, creating 50% shoot loss plots. Another randomly selected seven blocks were thinned to 20% ambient shoot density (80% shoot loss plots), and the remaining seven remained at ambient shoot density.

To conduct experimental trials, we affixed locally collected mesograzers to 10 cm pieces of monofilament (Fireline™; dia. 0.13 mm) tied near the top of 40 cm clear acrylic rods. After being tethered in the lab, each mesograzer was held in seawater overnight before being deployed to the center of a randomly chosen plot, 15 cm above the sediment surface, between 0800 – 1100 h the next morning. Trials lasted 24 h, at which time we retrieved acrylic rods and scored each individual as alive, eaten (fragments of the carapace remaining on the tether), missing, or molted (entire carapace remaining on the tether). We considered organisms that went missing to have been consumed by predators because no organisms tethered in predator-free controls at three sites (n = 20 mesograzers at Bodega Bay, Finland, and San Diego) fell off tethers after 48 h. Few animals molted on tethers, and any that did were removed from the analysis. Four trials of the experiment were conducted over a 7 – 10 day period at each site (N = 7 individuals per treatment per trial * 6 treatments * 4 trials = 168 organisms tethered per site).

Immediately after trials concluded at a site we sampled plots in which mesograzers were tethered for epibiont biomass and epifaunal biomass. Epibiont biomass represented the degree to which eelgrass shoots were colonized by epiphytic algae and sessile epifauna such as bryozoans; these organisms contribute to variability in structural complexity at very small scales. We used the biomass of mobile crustacean epifauna as a proxy for prey density. To quantify epibiont biomass, three shoots near the center of each plot were haphazardly selected and removed from the plot, and returned to the laboratory where all epibionts were scraped from shoots, dried, and weighed. Scraped shoots also were dried and weighed to calculate epibiont biomass per unit eelgrass biomass. Epifauna were sampled by placing a 500 m mesh bag with a 20 cm diameter opening over eelgrass in a haphazardly selected area of each plot. This method targets small mobile mesograzers, but not larger mesopredators. Captured organisms were removed from eelgrass blades in the laboratory, separated into crustaceans vs. others taxa (primarily gastropods), and weighed. Eelgrass collected in the bag was dried and weighed to standardize epifaunal biomass per unit eelgrass biomass.

Instruments: Experiments were conducted individually at each site, so instrumentation varied among sites. At each site, a balance was used to measure biomass of organisms collected in plots. In the field, sampling of organisms was performed by collecting organisms in mesh bags or by clipping shoots. Shoot counts were made using PVC or wire rings laid over plots.

Table 1. (A) Sites used in the tethering experiment, their locations, and principle investigators involved in the study. (B) Taxa used for the tethering experiment at each site.

Code	Site	Principle Investigator	Latitude	Longitude
BB	Bodega Bay, California, USA	J. Stachowicz	38.379	-123.053
CR	Posejarje, Adriatic Sea, Croatia	C. Kruschel	44.211	15.491
FI	Angso Island, Baltic Sea, Finland	C. Boström	60.108	21.711
FR	Bouzigues, Mediterranean Sea, France	F. Rossi	43.446	3.661
JN	Shinryu, Hokkaido, Japan	M. Nakaoka	43.052	144.842
JS	Akiwan Bay, Hiroshima, Japan	M. Hori	34.294	132.915
KO_A	Dong-dae Bay, Korea	K-S Lee	34.894	128.017
KO_B	Koje Bay, Korea	K-S Lee	34.800	128.583
MX	Punt Banda Estuary, Baja, Mexico	C. Hereu, P. Jorgensen	31.752	-116.626
NC	Back Sound, North Carolina, USA	J. Fodrie	34.671	-76.573
NI	Greyabbey, Irish Sea, Northern Ireland	N. O'Connor	54.519	5.562
OR	Sally's Bend, Oregon, USA	F. Nash	44.613	-124.013
QU	Point-Lebel, Quebec, Canada	M. Cusson	49.081	-68.311
SD	San Diego Bay, California, USA	K. Hovel	32.714	-117.171
SF	San Francisco Bay, California, USA	K. Boyer	37.940	-122.409
VA	Chesapeake Bay, Virginia, USA	E. Duffy	37.220	-37.254
WA	Willapa Bay, Washington, USA	J. Ruesink	46.497	-124.025

(B) Taxa used for the tethering experiment at each site.

Code	Site	Principle Investigator	Latitude	Longitude
BB	Bodega Bay, California, USA	J. Stachowicz	38.379	-123.053
CR	Posejarje, Adriatic Sea, Croatia	C. Kruschel	44.211	15.491
FI	Angso Island, Baltic Sea, Finland	C. Boström	60.108	21.711
FR	Bouzigues, Mediterranean Sea, France	F. Rossi	43.446	3.661
JN	Shinryu, Hokkaido, Japan	M. Nakaoka	43.052	144.842
JS	Akiwan Bay, Hiroshima, Japan	M. Hori	34.294	132.915
KO_A	Dong-dae Bay, Korea	K-S Lee	34.894	128.017
KO_B	Koje Bay, Korea	K-S Lee	34.800	128.583
MX	Punt Banda Estuary, Baja, Mexico	C. Hereu, P. Jorgensen	31.752	-116.626
NC	Back Sound, North Carolina, USA	J. Fodrie	34.671	-76.573
NI	Greyabbey, Irish Sea, Northern Ireland	N. O'Connor	54.519	5.562
OR	Sally's Bend, Oregon, USA	F. Nash	44.613	-124.013
QU	Point-Lebel, Quebec, Canada	M. Cusson	49.081	-68.311
SD	San Diego Bay, California, USA	K. Hovel	32.714	-117.171
SF	San Francisco Bay, California, USA	K. Boyer	37.940	-122.409
VA	Chesapeake Bay, Virginia, USA	E. Duffy	37.220	-37.254
WA	Willapa Bay, Washington, USA	J. Ruesink	46.497	-124.025

crustacean_survival.csv