FISHERY BULLETIN: VOL. 82, NO. 2 



period of maximum seasonal immigration for P. 

 aztecus as described by Baxter and Renfro (1967). 

 Accordingly, seven sets of samples were taken be- 

 tween 29 March and 23 July 1982. Each set was 

 obtained over a period of 3 d, and sets were taken 

 biweekly (29 March through 28 May) and monthly 

 (28 May through 23 July). Ordinarily, a set con- 

 tained 12 sample pairs that were subdivided to 

 sample the inner, middle, and outer marsh zones 

 equally, i.e., during each of three sampling days 

 four vegetated-nonvegetated sample pairs were 

 taken from a single zone. Sample sites within 

 zones were chosen randomly each month from 

 subunits in a grid superimposed on a map of the 

 area. The map and aerial photographs were used 

 to estimate percent coverage of vegetated and 

 nonvegetated habitats within different zones. 



A t-test of paired observations (Steel and Torrie 

 1960) provided the primary means for evaluating 

 differences in shrimp density between habitats. 

 Other analyses were performed using Pearson 

 product-moment correlations and ANOVAs across 

 sample sets, and Kendall's nonparametric concor- 

 dance tests (Tate and Clelland 1957) within sam- 

 ple sets. Analyses across sets incorporated an ele- 

 ment of temporal variability that was specifically 

 eliminated in analyses within sets. Data were log 

 transformed for ANOVAs to assure homogeneity 

 of variances. 



Procedures 



A drop sampler (Fig. 3) was designed to operate 



in the marsh from the bow of a skiff. The device 

 was an open-ended fiber glass cylinder, reinforced 

 on one end with galvanized metal, that enclosed 

 2.8 m 2 of marsh bottom. The sampler was deployed 

 endwise and pushed at least 15 cm into the sub- 

 strate to insure a good seal against leakage. After 

 marsh grass was removed, water was pumped 

 from the sampler and the enclosed bottom was 

 swept with dip nets to capture the entrapped or- 

 ganisms. The water and the contents of the dip 

 nets were placed into a 1 mm square mesh 

 plankton net with a removable cod end bag. When 

 all sample contents were washed, the cod end bag 

 was detached, labelled, and stored in a container 

 with Formalin 4 and Rose Bengal stain. 



Two identical sampling cylinders were used to 

 obtain sample pairs. Typically, the first sampler 

 was hoisted above the bow of the skiff and quietly 

 maneuvered into position over either vegetated or 

 barren substrate. The device was released and 

 allowed to free fall to the bottom. After disconnect- 

 ing the first sampler, the second sampler was 

 hoisted and the operation repeated in the opposing 

 habitat. The sequence of habitats was reversed 

 from pair to pair so that one would not continually 

 precede the other. Sample pairs were always 

 within two sample diameters of each other (3.6 m) 

 and care was taken to not disturb the site until the 

 second sampler was deployed. 



Within all samples, the water temperature, 



•"Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



FIGURE 3— A hand-operated drop 

 sampler used to estimate Penaeus az- 

 tecus densities in a Galveston West Bay 

 salt marsh. 



1.83 M DIA FIBER 

 GLASS CYLINDER 



METAL 

 SKIRT 



BOOM 



RELEASE LINE 





328 



CROSS SECTION 

 OF SAMPLER 



