FISHERY BULLETIN: VOL. 84, NO. 2 



The sample site was located 5 km southwest of the 

 E confirm River mouth (permanent station E-12 

 (Livingston 1975)). This site is characterized by 

 relatively uniform, dense stands of the seagrasses, 

 Thalassia testudinum and Syringodium filiforme, 

 with seasonal occurrence of red drift algae (mean 

 annual macrophyte biomass = 320 g dry wt/m 2 ; see 

 Zimmerman and Livingston 1979 for a description 

 of macrophytes). Station E-12 was polyhaline, with 

 salinities during collections ranging from 22 to 30 

 ppt (x = 27.0 ppt). Water temperatures ranged from 

 12.0° to 31.0°C (x = 19.9°). Depth varied from 1.6 

 to 2.1 m. Physical characteristics are summarized 

 in Table 1. 



Table 1.— Physical characteristics of the 

 sampling station for collection dates, 

 Apalachee Bay, FL. 



A 90 cm wide commercial Chesapeake Bay crab 

 scrape (Miller et al. 1980) was fitted with the cod 

 end of a 5 m otter trawl (6 mm mesh liner). The crab 

 scrape was towed at about 1.4 knots for 1 min (after 

 Greening and Livingston 1982; Leber 1983), yielding 

 a standardized tow of 42 m (mean of 10 preliminary 

 measured 1-min tows). A 42 m weighted line was 

 then used to standardize scrape tows during collec- 

 tions. A 5 m otter trawl (19 mm mesh wings, 6 mm 

 mesh liner in the cod end) was towed at the same 

 speed for 2 min (as in Livingston 1975, 1982; Hooks 

 et al. 1976; Heck 1977, 1979; Orth and Heck 1980; 

 Stoner 1980; Stoner and Livingston 1980; Dugan 

 and Livingston 1982; Dugan 1983), covering an 

 average measured distance of 84 m. Under tow, the 

 trawl mouth tickler chain fished a 2.1 m wide path 

 over the substratum (Leber, pers. obs.). Hence, each 

 individual trawl tow fished over 4.6 times the sub- 

 stratum surface area sampled by each tow of the 

 crab scrape (176 m 2 vs. 38 m 2 ). Because the scrape 

 collected larger amounts of dead vegetation, it was 

 logistically difficult to sample as much surface area 

 with it as was sampled by the trawl. 



Collections were made quarterly (January, April, 

 July, and October). On each sampling date eight 

 scrape and four trawl tows were taken (in the se- 

 quence two trawls, eight scrapes, two trawls) dur- 

 ing the day, and again beginning 1 h after dark. 

 Greening and Livingston (1982) determined that 

 eight 1-min scrapes were sufficient for sampling 

 >95% of the species of macroinvertebrates at our 

 sample site in Apalachee Bay. Because each scrape 

 was towed for only half the 2-min towing time used 

 for each trawl (scrape tows lasting longer than 1 min 

 often resulted in clogging the net with red drift 

 algae), only four trawls were taken during each sam- 

 pling period. Thus, the combined length of the eight 

 scrape tows (8 x 42 m = 336 m) matched that of 

 the four trawl tows. All samples were collected from 

 a 0.25 km 2 area immediately south of the station 

 marker. Replicate tows were taken along transects 

 spaced at least 30 m apart to prevent overlapping 

 samples. 



Organisms were preserved in 10% Formalin 4 (buf- 

 fered with seawater) in the field, then identified, 

 counted, and measured in the laboratory. A two-way, 

 Model II, factorial ANOVA design for unequal but 

 proportional cell sizes (Sokal and Rohlf 1969) was 

 used to compare mean numbers of species and in- 

 dividuals of each taxon group in scrape vs. trawl 

 (Factor 1) and day vs. night (Factor 2) samples. 

 Log 10 transformations were used where F-max 

 tests indicated heterogeneity of variance Rather 

 than extrapolating our data to numbers per unit 

 area, we compared the collections made with these 

 two gears using absolute numbers per tow in our 

 calculations (which are biased in favor of the trawl 

 by a factor of 4.6). We used these absolute abun- 

 dances because 1) we wanted a strongly conservative 

 test of our premise that the scrape is the more ef- 

 fective of these two sample gears in vegetated 

 aquatic habitats, and 2) we believe that extrapola- 

 tions of semiquantitative data to abundances per unit 

 area yield highly unrealistic results, which may be 

 misinterpreted by readers as accurate densities (cf. 

 Howard 1984, who determined that a towed beam 

 trawl was only 4.7% efficient in estimating densities 

 of shrimp in an Australian seagrass meadow). 



RESULTS 



Factor 1: Trawl vs. Scrape 



Although the surface area sampled by the otter 



"Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



444 



