FISHERY BULLETIN: VOL. 85, NO. 1 



(recently molted) were used in the experiments. 

 Tethering of crabs was accomplished by tieing one 

 end of a 1 m long piece of monofilament fishing line 

 around the width of the body and securing the loop 

 with "Super" glue (cyanoacrylate) to the top of the 

 carapace. The other end of the line was tied to a J- 

 shaped, heavy piece of wire (or stake) which was 

 pushed into the sediment in the chosen seagrass or 

 sand location. The super glue ensures that crabs do 

 not escape and that a piece of the carapace is left 

 on the line as evidence if predation does occur. The 

 carapace width (CW) of all crabs was measured 

 before placement in the field. For blue crabs larger 

 than 40 mm CW, a 20-lb test steel leader was at- 

 tached to the monofilament loop around the crab to 

 prevent the cutting of the tether by the crabs' claws. 

 Tethering techniques measure relative rates of pre- 

 dation and are used for comparison of mortality 

 among sites. It is not intended to measure absolute 

 rates of predation in any single habitat. Heck and 

 Thoman (1981) provided an additional description 

 of the tethering procedure. 



A single blue crab was tethered to an individual 

 stake, and three to four stakes were placed in each 

 plant density and in unvegetated sand patches for 

 each 24-h trial. The tethered crabs were left at the 

 site for 24 h (-t-/-l h), recovered, and predation 

 losses scored. Twenty trials, utilizing a total of 218 

 crabs, were conducted from 15 July through 7 

 October. 



The density of the seagrass was determined fre- 

 quently during the study period by measuring dry 

 weight biomass of the grass removed from 0.062 

 m^ plots. Four samples with three replicates for 

 each sample at each density were taken, and dry 

 weights measured after drying at 100 °C. 



RESULTS 



Vegetation clearly provides cover from predators 

 for blue crabs (Fig. 1) as predation was always more 

 intense in unvegetated sand patches than in sea- 

 grass. Relative rates of predation on tethered crabs 

 on sand ranged from 24% to a high of 74% eaten 

 per day. A 3-way contingency table analysis (sur- 

 vival X density x date) found significant interac- 

 tions (P < 0.01) between crab survival and density 

 of vegetation. Differences in predation rates among 

 time periods were not statistically significant, al- 

 though predation rates dropped steadily on sand 

 after the middle of August and no predation was 

 recorded in vegetation in October (Fig. 1). The in- 

 fluence of body size (CW) of crabs (Fig. 2) on risk 

 to predation was also tested in a Kolmogorov- 



FlGURE 1.— Percent juvenile blue crabs eaten in sand (S) or vegeta- 

 tion (V), July to October 1985. Time period is broken into 2-3 wk 

 periods. 



Smirnov test and found not significant (P > 0.05). 



Predation rate varied among densities of eelgrass 

 (Fig. 3). Medium density seagrass provided the best 

 refuge from predation with only 9% eaten per day 

 (A'' = 45). A mean of over 19% per day was eaten 

 in low-density (N = 47) and high-density {N = 44) 

 grass sites. A Dunn's Multiple Comparison test 

 (Hollander and Wolfe 1973) was used to analyze the 

 predation-vegetation density data from July through 

 September, excluding October because no predation 

 occurred in eelgrass during that month. Predation 

 rates in low and high densities were found to be sig- 

 nificantly greater (P < 0.05) than in medium-density 

 eelgrass. 



Eelgrass biomass in low, medium, and high den- 

 sity 0.062 m^ plots (Table 1) was found to be sig- 

 nificantly different in a one-way analysis of variance 

 (P < 0.001). Scheffe contrasts found that the mean 

 dry weight of medium-density plots was significantly 

 higher than low-density and significantly lower than 

 high-density eelgrass plots. 



Table 1.— Mean dry weights (g/0.062 

 m^) of vegetation from experimental 

 plots. ** Significantly different at the 

 P < 0.01 level. 



These data confirm results from other experimen- 

 tal studies of predation on decapod crustaceans 

 (Heck and Thoman 1981; Orth and van Montfrans 



54 



