28.5°N) at depths ranging from 4 to 20 m. Dates 

 of the four seasonal cruises were as follows: sum- 

 mer, 15 July-20 September 1980; spring, 28 April- 

 6 June 1981; winter, 7-29 January 1982; fall, 14 

 October-7 December 1982. Sampling locations were 

 selected by means of a stratified random sampling 

 design (Grosslein 1969). Thirteen strata were estab- 

 lished at depths of 4-12 m from Cape Fear to the 

 mouth of the St. John's River, FL (lat. 30.4°N), 

 while south of this point to Cape Canaveral, an addi- 

 tional five strata were delineated in depths rang- 

 ing from 8 to 20 m with the 5.6 km territorial sea 

 line as the offshore boundary. A change in defini- 

 tion of strata off Florida was necessary because of 

 the steepness of the nearshore continental shelf in 

 the area south of the St. John's River. Areal extent 

 of strata ranged from 7,486 to 31,661 ha (x = 

 17,323 ha). 



At each randomly selected site within a stratum, 

 paired tows were made for 20 minutes at a speed 

 of 4.4 km/h using four-seam Gulf of Mexico shrimp 

 trawl nets. Outriggers enabled two nets to be fished 

 simultaneously from the same vessel. The trawl 

 nets, which were attached to 1.5 x 0.8 m wood and 

 chain doors, had headrope lengths of 12.8 m, foot- 

 rope lengths of 15.8 m and stretch-mesh sizes of 5.1 

 cm in the wings, 4.4 cm in the body, and 4.1 cm in 

 the cod end. Tickler chains were attached to each 

 door and adjusted to drag at a distance of 0.6 m in 

 front of the nets. Tows were confined to daylight 

 hours (1 hour after sunrise to 1 hour before sunset) 

 to eliminate day-night changes in availability and 

 vulnerability of organisms to the gear. Sampling 

 effort changed from cruise to cruise owing to con- 

 straints of funding and vessel availability (Fig. 1). 

 Collections of decapod and stomatopod crustaceans 

 taken during tows in which a net was damaged, 

 failed to reach bottom, or became twisted were con- 

 sidered unsuccessful and were not included in anal- 

 yses. Bottom temperature was measured following 

 each tow with a stem thermometer mounted within 

 a Van Dohrn bottle. 



Decapod and stomatopod crustaceans collected by 

 each net during a tow were identified to species, 

 counted, and weighed to the nearest gram, and 

 either the entire catch or a random subsample of 

 at least 30 individuals was measured to the nearest 

 mm. Measurements included total length (tip of 

 rostrum to tip of telson) for shrimp [carapace length 

 (tip of rostrum to posterior margin of carapace) for 

 majid crabs, and carapace width (measured between 

 the posteriormost lateral spines) for other crabs]. 

 Sex was recorded for brachyuran species. Although 

 the contents of each net were processed indepen- 



FISHERY BULLETIN: VOL. 87, NO. 1 



dently, the catch per standard tow is defined as the 

 combined catch from both trawl nets fished simul- 

 taneously at each randomly chosen site within a 

 stratum. 



Data Analysis 



The stratified mean catch per tow and its stan- 

 dard error were calculated by the expressions (Poole 

 1974): 



VST 



L 



h.i N 



NkVh 



SE-^ = I ^W,- 





1 - 



Hi 



where ysr = stratified mean catch/tow 



2/;, = mean catch/tow in numbers or weight 



in the h stratum 

 N^ = number of possible sampling units 



in the h stratum 

 N = total number of possible sampling 



units over all strata 

 SEy^^ = standard error of the stratified mean 



catch/tow 

 W, = N,IN 



Si,~ = sample variance of the h stratum 

 ri), = number of trawl tows madp m the 



h stratum. 



The effective degrees of freedom for the calculation 

 of confidence limits follows the methodology of 

 Cochran (1977). 



Minimum biomass and density estimates for trawl- 

 caught groundfish were calculated from computa- 

 tions of the area swept by the trawl gear. The ap- 

 proximation of the area swept (a) was derived from 

 the following equation (Roe 1969): 



_ K X M X (0.6 H) 

 10,000 m^/ha 



where a = swept area in hectares 



K = speed in meters per hour 



M = time in hours fished 



H = headrope length in meters. 



The result (1.12 ha) was multiplied by two since the 

 standard unit of effort was two nets fished simul- 

 taneously. Thus, a standard trawl station sampled 

 2.24 ha. 



156 



