of marked shrimp was random (one-sample runs test, 

 P = 0.960; Siegel 1956). ^ 



The population estimate of 207,786 shrimp deter- 

 mined from mark-recapture data compared favorably 

 with the results of a concurrent drop sampler ex- 

 periment (Ikble 1). Shrimp densities were obtained 

 using a 2.8 m^ drop sampler at high tida Detailed 

 methodology has been described by Zimmerman et 

 al. (1984). Drop samples were taken in two sets, four 

 pairs each, in vegetated and nonvegetated areas, 

 divided between the south and north ends of the 

 bayou. Vegetated habitat was sampled along the 

 bayou margins, while nonvegetated area sampling 

 was in open waters of the bayou. Numbers of shrimp 

 within the sampler were extrapolated to represent 

 the shrimp population in the vegetated, nonvege- 

 tated, and total areas of Sydnor Bayou. Confidence 

 intervals for the drop sampler were much wider than 

 those for Petersen estimate because drop sampler 

 estimates were based on a small number of samples. 

 The drop sampler estimate for 36 ha was higher by 

 about 92,000 shrimp. One reason for this difference 

 is that the mark-recapture estimate reflects only that 

 part of the population >40 mm TL, while the drop 

 sampler measures density of small (<40 mm) shrimp 

 more effectively, and these small shrimp are included 

 in the estimate (Ikble 2). We calculated the drop sam- 

 pler population estimate using only shrimp larger 

 than 40 mm TL (Tkble 1). A chi-square test shows 

 a significant difference between the drop sampler 

 and mark- re capture size-frequency samples, cate- 

 gories 41-50 mm and higher (x^ = 109.45, df = 6, 

 P very small). The high chi-square value is due main- 

 ly to the greater number of 41-50 mm shrimp and 

 the lower number of larger shrimp (81-90 mm), which 

 may avoid the sampler, in the drop sampla Length- 

 frequency composition of the drop sampler catch in- 

 dicates that 23% of the 103 shrimp taken were 

 smaller than 40 mm TL, while no shrimp smaller 

 than 40 mm were captured by the otter trawl. 



1970-71 Population Estimates 



Our methodology for conducting a Petersen single 

 census mark-recapture experiment with juvenile 

 brown shrimp was developed during June and July 

 1971 studies of five Tbxas coastal ponds (Fig. 1). All 

 ponds ranged from 0.3 to 0.9 m in depth during a 

 normal summer tidal (ycle Cow Trap 1 and 2 had 

 considerable emergent vegetation along their shore- 

 lines and were part of a large marsh complex. Ex- 

 tensive flooding of the marsh surrounding these 

 ponds at flood tide greatly increased the area ac- 

 cessible to shrimp, but this shallow, vegetated area 



Table 1. — Sydnor Bayou brown shrimp population 

 estimates determined by mark-recapture and drop sampler 

 methods, June 1983. 



'Estimate of shrimp population >40 mm TL. 



Table 2. — Length-frequency composition 

 of Sydnor Bayou brown shrimp samples 

 taken with the otter trawl (N = 8,197) and 

 drop sampler (W = 83), 1-3 June 1983. 



could not be sampled. Shrimp could move from pond 

 to pond via flooded marsh and ditches, rendering 

 block nets ineffective Evidence of this movement 

 was the netting of marked shrimp released in Cow 

 Trap 1 and recaptured in Cow Trap 2. These 

 problems precluded reasonable population estimates 

 for the Cow Trap ponds, and large marsh complexes 

 were avoided for future studies of this typa 



Mud Lake, Carancahua, and Mushroom had 

 generally well-defined shorelines, even during flood 

 tide, and were not contiguous with other ponds or 

 ditches. Mark-recapture methods were essentially 

 the same as described for the 1983 study. Marking 

 and holding operations were conducted on a portable 

 barge rather than from shore (Emiliani and Marullo 

 1973). Population estimates determined by Bailey's 

 (1951) formula ranged from 7,490 to 17,119 brown 

 shrimp per hectare (Ikble 3). The lowest estimate 

 was recorded in Mud Lake, where the highest 

 percentage of total catch was <40 mm TL, while the 

 highest estimate was for Carancahua. The density 

 in Mushroom was close to that in Carancahua. 



Although marking methods differed, a 1970 mark- 



680 



