N = M 



(C + 1) 



R + 1 



N=122 



where M = number of marked shrimp released, 

 corrected for marking mortahty, 

 C = number of shrimp examined for marks, 

 R = number of recaptured marked shrimp 

 in the sample 



The 95% confidence limits for the true population 

 were estimated using the standard error of the large 

 sample variance formula (Bailey 1951) 



V(N) = 



M^jC + 1){C - R) 

 (R + 1)2 (R + 2) 



Application of this method is justified under the 

 following conditions (Ricker 1975): 



1) Marked shrimp suffer the same natural mor- 

 tality as unmarked. 



2) Marked and unmarked shrimp are equally 

 vulnerable to fishing. 



3) Marked shrimp do not lose their mark. 



4) Marked shrim.p become randomly distributed 

 among unmarked. 



5) All marks are recognized and reported on 

 recovery. 



6) There is not emigration or immigration occur- 

 ring in the catchable population. 



Results and Discussion 



Overall marking mortality was 9%. One cage had 

 unusually high mortality. Nineteen of 25 marked 

 shrimp were alive at the end of 24 h, and the only 

 evidence of the other 6 marked shrimp was pieces 

 of exoskeleton. They apparently molted and were 

 cannibalized. Holt (1982) suggested that the condi- 

 tion of shrimp prior to tagging dictates the survival 

 of the tagged animals. When stressed animals were 

 tagged, mortality more than doubled. Howe and 

 Hoyt (1982) hypothesized that tags and marks may 

 indirectly cause mortality by attracting predators. 

 Farmer and Al-Attar (1979) found shrimp marked 

 with subcutaneous pigment suffered high mortality 

 (compared with controls) only when held with un- 

 marked conspecifics. Clark and Caillouet (1973), 

 however, found negligible marking mortality in a 

 mark-recapture experiment with white shrimp, R 

 setiferus, when 50 marked and 50 unmarked con- 

 trol shrimp were held in a large pen in a pond rather 

 than in several small cages. Costello and Allen (1962) 

 stated that stained shrimp may be expected to sur- 



1 

 20 



40 60 80 100 



TOTAL LENGTH (MM) 



120 



Figure 2.— Length-frequency distribution of brown shrimp in Syd- 

 nor Bayou, June 1983: A) representative sample of shrimp col- 

 lected during marking; B) unmarked shrimp caught during recap- 

 ture operations; and C) marked shrimp caught during recapture 

 operations. 



vive at essentially the same rate as unmarked 

 shrimp, regardless of presence of predators, lb avoid 

 overestimating marking mortality, we did not include 

 the counts in the high cage in the calculation. The 

 resulting 4% (3 dead marked shrimp out of 75) was 

 similar to the marking mortalities of past studies in 

 Sydnor Bayou, Mud Lake, and Mushroom (Welker 

 and Baxter fn. 1). 



1983 Population Estimate 



A total of 223 marked shrimp were among 12,304 

 shrimp caught in 94 recapture tows. Tides during 

 the recovery period were low in the morning, ap- 

 proaching high tide in the afternoon. Areas along 

 the shore and the south end of the bayou were 

 shallow for trawling in the mornings, but could be 

 adequately sampled in the afternoon. Distribution 



679 



