Broadhurst et al.; Flow-related effects on prawn-trawl codends 



scribed above but included 

 composite square-mesh panels 

 made of 60-mm and 40-mm 

 netting (3-ply and 48-ply UV- 

 stabilized high-density poly- 

 ethylene twine, respectively) 

 cut on the bar and inserted 

 into the tops of the anterior 

 sections (termed the 100 panel 

 and 200 panel codends) (for 

 details see Fig. IC and Broad- 

 hurst and Kennelly, 1996). Be- 

 cause the current meter (and 

 not its stanchion) was inserted 

 only 5 cm into the tops of the 

 codends (Fig. IC), it was not 

 necessary to cut the meshes of 

 these codends. Instead, four 

 positions were labeled with a 

 permanent marker at 2200 

 mm, 1720 mm, 1490 mm, and 

 1130 mm forward from the end 

 of the codends (Fig. IC). 



Experimental procedure 



Thirty five rubber balloons 

 were each filled with 2 liters 

 of water, providing a total 

 mass (in air) of 70 kg. These 

 balloons were used to simulate 

 masses of catch in the codends. 

 In each experiment, the two 

 codends were tested alter- 

 nately. The particular codend 

 to be examined was attached 

 to the trawl and loaded ini- 

 tially with 15 balloons (i.e. 30 

 kg). The hydraulic pumps in 

 the flume tank were activated 

 and adjusted to produce a flow 

 of 1.2 m/s (the standard tow- 

 ing speed during commercial 

 operations). After a stabilizing 

 period of 10 minutes, the stan- 

 chion containing the current 

 meter was alternately lowered 



into each of the predetermined positions in the 

 codends (three in experiment 1 and four in experi- 

 ment 2). After a further stabilizing period of one 

 minute at each position, the current meter was 

 switched on and left for a period of one minute, dur- 

 ing which the flow of water immediately anterior to 

 the current meter was recorded at one second inter- 

 vals and the data were transmitted to the computer. 



2,200 mm 



2 kg 

 Balloons 



2,200 mm 



Figure 1 



Diagrammatic representation of (A) the current meter and its stanchion, (B) 100 and 

 200 commercial codends and the three positions of water flow readings and (C) 100 

 and 200 panel codends and the four positions of water-flow readings. T = transver- 

 sals; and N = normals. 



The mean flow rate from each minute of recording 

 was calculated from these data and used in subse- 

 quent analyses. After each reading, the current meter 

 was moved to the next position and the procedure 

 repeated. After six replicate readings were collected 

 for each position, the flow of water in the flume tank 

 was reduced to approx. 0.5 m/s and additional bal- 

 loons were added to the trawl to simulate an increase 



