FISHERY BULLETIN; VOL. 87. NO. 4, 1989 



yielded e^ = 0.18 and w = 0.54 (Table 2), thus, w,. 

 = 0.33. According to these results, the purse 

 seine seems comparably efficient for both species 

 (12% and 18%), but the avoidance and the es- 

 capement rates are very different. 



Species and Size Selectivity 



The total catches by the purse seine (25 sets) 



Table 8. — Total catches in number by purse seine (PS) and 

 beach seine (BS) during the whole experiment. The mean 

 ratio of catches per set by both gears is calculated with the 

 coefficient of variation (C.V.) expressed as a percentage. 



Species 



PS 



BS Ratio C.V. 



Genes nigh 167 8,591 0.0039 



Pomadasys jubelini 6 143 0.0084 



Tylochromis jentinki 22 366 0.0120 



Chrysichthys auratus 139 2,232 0.0125 



Chrysichthys maurus 26 384 0.0135 



Chrysichthys nigrodlgitatus 9 96 0.0187 



Caranx hippos 2 17 0.0235 



Elopslacerta 135 1,049 0.0257 



Anus latiscutatus 6 32 0.0375 



Ethmalosa fimbhata 1,958 8,816 0.0444 



Trachinotus teraia 1 1 44 0.0500 



Sarotherodon melanotheron 27 81 0.0667 



Callinectes amnicola 96 137 0.1401 



Cithahchthys stampflii 61 75 0.1627 



Tilapia guineensis 362 437 0.1657 



Penaeus notialis 1,234 32 7.7125 



Total 2,973 22,963 0.0259 



130 

 101 



32 

 127 

 120 

 127 

 132 



39 

 173 



37 

 148 

 156 

 261 

 145 



18 



92 



and by the beach seine (5 sets) are summed in 

 Table 8, along with the ratio of these values. 

 Important difference appear between the ratios 

 for the 16 species listed. They range between 

 0.003 and 0.17; this will be termed "species selec- 

 tivity". 



The same ratio was also computed by size 

 group for seven species and analyzed using a 

 Kruskall-Wallis one-way analysis of variance 

 (Table 9). A large size effect, termed here "size 

 selectivity", appears for T. guineensis and is 

 Hkely for Gen-es nigri, but is not significant for 

 the other species. 



DISCUSSION 



The validity of the different results depends to 

 a large extent on the robustness to departures 

 from the underlying assumptions: no marking 

 and holding stress, no mortalities, and no en- 

 closure effect (e.g., accessible stock may differ 

 from the standing stock if fish are crowded along 

 the enclosure). 



Escapement 



For the Chrysichthys spp., the estimates of 

 the retention rates are consistent with estimates 

 of catchability and seem valid. On the other 

 hand, for T. guineensis, the retention rates lead 

 to unexpected results on three occasions (Wi > 

 W2, Table 4). Marked-1 fishes escaped less than 



Table 9. — Comparison of the catches per set by both gears, by size group, using a 

 one-way analysis of variance by ranks (Kruskall-Wallis test). Critical value of x^0 95 

 is 7.81. 



Size group 



Species 



1 



.2 



\ obs 



Tilapia guineensis 

 Genes nigri 

 Elops lacerta 

 Ethmalosa fimbhata 

 Chrysichthys auratus 

 Chrysichthys maurus 



R 1.127 0.804 0.556 0.154 12.78 



C.V. 35 23 48 32 



R 0.012 0.023 0.036 0.102 6.15 



C.V. 105 130 99 54 



R 0.144 0.092 0.113 0.244 0.09 



C.V. 18 181 87 32 



R 0.202 0.454 0.376 1.091 2.38 



C.V. 58 1,289 166 201 



R 0.065 0.065 0.030 0.077 3.44 



C.V. 130 164 181 235 



R 0.162 0.040 0.029 5.82 



C.V. 175 294 84 



Chrysichthys nigrodlgitatus R 0.087 0.187 0.080 2.30 



C.V. 57 140 163 



918 



