Fishery Bulletin 97(1), 1999 



1.2 



Position no. 1 



09 



08 



12 



B Position no. 2 



•o 



c 

 o 



B ^2-, 



C Position no. 3 



significant increase in the flow 

 of water at position no. 4 in the 

 200 panel codend with 30 kg 

 and a similar result (though 

 not statistically significant) for 

 50 kg (Fig. 3D; Table 2). This 

 anomaly may be explained by 

 the fact that the balloons in 

 the 200 panel codend (like 

 those in the 200 commercial 

 codend) were orientated evenly 

 across the surface area of the 

 posterior section, increasing 

 its diameter and decreasing 

 the angle of incidence of its 

 netting as it led into the ante- 

 rior section of the codend at 

 position no. 4. Although the 

 current meter was located im- 

 mediately under the compos- 

 ite square-mesh panel, at this 

 position it was effectively 

 aligned slightly above the an- 

 terior section of the codend 

 and may have been influenced 

 by the current outside the 

 trawl, negating some of the 

 flow-related effects of codend 

 mesh circumference. 



With the exception of this 

 latter result, the measured 

 reductions in flow at most po- 

 sitions in the 200 codends (cor- 

 responding to increases in 

 water displacement forwards) 

 support the hypothesis that an 

 increase in the circumference 

 of meshes in the codend con- 

 tributes towards the escape of 

 small fish (between 5 and 20 

 cm) through the composite 

 square-mesh panel (see Broad- 

 hurst and Kennelly, 1996; 

 1997). The size of these fish 

 suggests that they are using 



anaerobic muscle power to maintain position in the 

 moving trawl (1.2 m/s) and are fatigued when they 

 enter the codend (see Beamish, 1978; Wardle, 1989). 

 A relatively small increase in the displacement of 

 water forwards (e.g. 0.203 m/s at position no. 3 in 

 the 200 commercial codend or 0.1 m/s at position no. 

 3 in the 200 panel codend I may be sufficient 1) to 

 assist small fish to swim forwards and out through 

 the square meshes in the panel and [or| 2) to enable 

 them to reduce their tail-beat frequencies and main- 



Water speed 

 in flume tank 



lllxili 



Water speed 

 in flume tank 



llJill 



Water speed 

 in flume tank 



LiJi 



1.2 T^^ 



D^ositiormo. 4 Water speed 



in flume tank 



1,1 - 



08 



iJilU 



100 

 Panel 



200 



Panel 



100 

 Panel 



200 

 Panel 



100 

 Panel 



200 

 Panel 



30 kg 



50 kg 



70 kg 



Figure 3 



Differences in mean flow rates ±SE between the 100 and 200 panel codends tested in 

 experiment 2 for each position of the current meter and for different weights {< and > 

 indicate direction of differences in SNK tests of means). 



tain their position in the codend for a longer period, 

 increasing their chances of random escape through 

 the panel; and (or) 3) to stimulate their lateral line 

 receptors and thus their overall escape. 



Without direct observations of fish swimming in 

 the codend, it is difficult to determine their specific 

 behavior during escape. Whatever their actual es- 

 cape mechanism, however, the results obtained in 

 this study provide important information for the sub- 

 sequent design and location of BRDs like the com- 



