654 
Fishery Bulletin 95(4), 1997 
several trawl fisheries in the North Atlantic involves 
inserting large panels of square-mesh in codends 
(Robertson and Stewart, 1988; Carr, 1989; Briggs, 
1992; Isaksen and Valdermarsen 2 ; Suuronen 3 ). 
These studies have shown that square-mesh panels 
often reduce the bycatch of juvenile roundfish while 
retaining a large proportion of the targeted catch. In 
previous experiments (Broadhurst and Kennelly, 
1994, 1995, 1996; Broadhurst et al., 1996), we have 
shown that relatively small panels of square-mesh, 
inserted into the top anterior sections of penaeid 
prawn-trawl codends, allowed large numbers of small 
fish to escape without any losses of prawns. In these 
experiments, the majority of fish were thought to 
have been herded together in the narrow anterior 
section of the codend, immediately in front of the 
catch (see also Wardle, 1983). This concentration of 
fish was thought to upset the balance of the school 
and to initiate a response in the fish to escape by 
swimming towards the sides and top of the net and 
out through the open square-meshes. In addition, we 
showed that codend circumference and differences 
in hydrodynamic pressure had significant effects on 
the rates of movement of these fish through the 
square-mesh panel. The reaction of prawns to these 
stimuli was considered to be fairly limited, given their 
inability to maintain an escape response to trawls 
(see Lochhead, 1961; Main and Sangster, 1985). 
In a recent experiment ( Broadhurst and Kennelly, 
1996) in one location in NSW, we tested a new de- 
sign of codend, comprising composite panels of 
square-shaped mesh (referred to as the composite- 
panel codend), designed for and located in the codend, 
to take advantage of the theory discussed above. The 
results showed that this design was effective in re- 
ducing up to 40% of the total unwanted bycatch and 
2 Isaksen, B., and J. W. Valdemarsen. 1986. Selectivity experi- 
ments with square mesh codends in bottom trawl. Int. Coun. 
Explor. Sea council meeting 1986/B: 28, 18 p. 
3 Suuronen, P. 1990. Preliminary trials with a square mesh 
codend in herring trawls. Int. Coun. Explor. Sea, council meet- 
ing 1990/B: 28, 14 p. 
up to 70% of the numbers of juveniles of commer- 
cially important species with no significant reduc- 
tion in the catches of prawns and other commercially 
important species. Although not validated statistically, 
there was also some evidence to suggest that the trawls 
with the composite square-mesh panel retained, on 
average, slightly more prawns than a conventional 
trawl (means increased by up to 3%). This latter re- 
sult, in particular, led numerous local fishermen to in- 
stall the composite-panel voluntarily in their trawls and 
use it as part of normal commercial operations. 
To assess the performance of this design through- 
out the full geographic range of this fishery (encom- 
passing the range of fishing conditions and catches) 
and to promote its voluntary acceptance, our specific 
goals in the present study were to investigate the 
effectiveness of the composite-panel under normal 
commercial operations at four major ports along the 
NSW coast in 1) reducing unwanted bycatch, 2) main- 
taining catches of commercially important byproduct, 
and 3) increasing catches of prawns. 
Materials and methods 
This study was performed between December 1995 
and February 1996 with four commercial vessels (see 
Table 1 for details) on prawn-trawl grounds offshore 
from four ports (Port Stephens, Southwest Rocks, 
Yamba, and Ballina) in New South Wales, Australia 
(Fig. 1). Each vessel was rigged with three Florida 
flyers (mesh size=42 mm) in a standard triple gear 
configuration (see Kennelly et al., 1993 for details), 
towed at 2.5 knots. Each of the identical outside nets 
on each vessel were rigged with zippers (no. 10 ny- 
lon open-ended auto-lock plastic slides) to facilitate 
removal and attachment of the codends. Because each 
of the middle nets were not rigged in exactly the same 
way as the outside nets, their catches were excluded 
from any analysis. 
The codends used in the study measured 58 meshes 
long (2.3 m) and were constructed from 40-mm mesh 
Table 1 
Summary of vessels, trawl headline lengths, and depths trawled for each of the four ports. 
Trawl headline 
Port Vessel and (length in m) length for each net (m) Depth trawled (m) 
Port Stephens 
Fairwind 
(16) 
16.45 
75-88 
Southwest Rocks 
Shelley-Anne 
(13.7) 
10.97 
47-53 
Yamba 
L- Margo 
(15.93) 
12.8 
20-49 
Ballina 
New Avalon 
(18.5) 
14.63 
29-55 
