CHARLES-DOMINIQUE; CATCH EFFICIENCIES OF PURSE AND BEACH SEINES 



ered by the gears, p„ = a/A; therefore, accord- 

 ing to Equation (I), e = q • A/a. 



Catchability q was estimated by using two 

 partially independent methods (see notations on 

 Table 1). In the first method (estimation "from 

 captures"), catchability was estimated for each 

 purse seining operation by the equation: q, = 

 Cj/Ni {i = 1...5), where c, represents captures 

 at set ; and N,, the standing stock in the en- 

 closure just before set /. When escapement did 

 not occur during the first period (Wi = W-?), the 

 standing stock after last purse seine set (A^^) was 

 estimated by the ratio of the beach seine cap- 

 tures (C) to the retention rate W. The previous 

 values of N, were then calculated backward (A^, 

 = A^, = i -I- c,, i = 5...1), and the q, values 

 follow. 



In the second method (estimation "from recap- 

 tures"), catchability was estimated from the re- 

 captures of marked-1 fish only when escapement 

 did not occur during the first period iW^ = W-t). 

 In this case, q, = ri,/Mi,, where M\, is the stand- 

 ing stock of marked-1 fish just before the set i. 

 Ml, was calculated by successive subtraction of 

 type-1 recaptures (Afi.,i = Mu = ru). 



The two methods of measuring q were com- 

 pared using a distribution-free test (Wilcoxon's 

 signed rank test, see Dagnelie 1975), which ap- 

 plies to paired data samples. The experimental 

 design described above is summarized in Table 

 3. 



Both methods rely on some underlying as- 

 sumptions: A) no mortality of marked fish 

 occurs; B) marked and unmarked fish have the 

 same probability of escaping from the en- 

 closure; C) the efficiency of the purse seine is 

 equal for marked and unmarked fish; and D) all 

 fish present in the enclosure have an equal prob- 

 ability of being caught. The method from cap- 

 tures relies on assumptions A), C), and D); the 

 method from recaptures relies on A), B), and D). 



RESULTS 



Tagging and Holding Tolerance 



No mortality of marked fish w-as observed dur- 

 ing our experiment. This included the holding- 

 period in the floating cages as well as the fishing 

 period (no dead marked fish were recovered in 

 the seines). During the preliminary tests, how- 

 ever, marked fish of less robust species (Eth- 

 malosa fimbriata and Eucinostomus mela- 

 nopterus) were found dead in both the cages and 

 the fishing nets. 



Retention Rates 



Purse seine captures and recaptures observed 

 in the 15 independent mark-recapture experi- 

 ments are listed in Table 2. The mean retention 

 rates and the coefficient of variation were calcu- 

 lated for seven fish species and for the portunid 

 crab Callinectes amnicola. The retention rate 

 ranged from 0.47 to 0.79 for fish and was 0. 10 for 

 C. amnicola, w'hich escaped in large numbers 

 probably by burying itself. No fish were ob- 

 served jumping over the purse seine net. There- 

 fore, escapement appeared to be due to fish go- 

 ing under the lead-line. 



Retention rates were estimated by size group 

 for the two principal species, Tilapia guineensis 

 and Chrysichthys spp. (grouping the three 

 species, C. maurus, C. auratus. and C. iiigro- 

 digitatus). No difference was found between the 

 size groups using a one-way analysis of variance 

 by ranks (Ki'uskall-Wallis test, see Table 5). 



The beach seine retention rates, Wi and Wo, 

 were calculated and their equality tested for the 

 two principal species Hsted above (Table 4). For 

 Chrysichthys spp., Wi was always less than or 

 equal to W-z- The mean retention rate (W) was 

 0.53. No size effect was found in the analysis of 

 variance by size gi'oup (Table 5). 



For T. guineeyisis, Wi was less than or equal 

 to W-, in only two experiments. The mean reten- 

 tion rate was 0.35. In the three other experi- 

 ments, the unexpected result of Wi being 

 gi'eater than W-z was found. This point will be 

 discussed later. Again, no size effect w-as noticed 

 (Table 5). 



Table 4. — Estimation of the beach seine retention rates W, 

 and W2 (number of recaptured over number of released fish). 

 If Wi and IV2 do not differ significantly in one experiment (x^ 

 test for the difference of two proportions, P = 0.05), the mean 

 value W is then calculated. Parentheses mean a departure 

 from the limit of application conditions of the test. 



Retention 

 rates 



Capture dates (Oct. 1984) 



8 



12 



13 



16 Mean 



Chrysichthys spp. 



W, 7/15 57/78 28/107 78/141 31/106 



W2 21/36 9/13 16/27 17/27 3/11 



X^ 0.58 (0.36) 11.5 0.41 (0.13) 



W 



0.53 



0.55 0.73 0.57 0.29 



Tilapia guineensis 



Wi 45/102 24/50 19/26 19/27 14/53 



W2 38/90 28/92 27/102 13/36 8/27 



0.07 4.31 19.5 7.2 0.09 



0.43 0.28 0.35 



X 

 W 



915 



