210 
Fishery Bulletin 95(2), 1997 
In estuarine prawn-trawl fisheries in NSW, many 
of the individual fish in bycatch are larger than the 
targeted prawns and include organisms such as jel- 
lyfish or jelly “blubber” — Catostylus spp. For the past 
30 years, many of the estuarine prawn-trawlers in 
NSW have routinely used a BRD designed specifi- 
cally to exclude these individuals. Commonly called 
“blubber-chutes,” these BRD’s consist of a funnel of 
soft mesh inserted into the aft belly of the trawl. 
Organisms larger than the mesh in the funnel are 
guided through an opening in the top of the trawl, 
while prawns and smaller individuals pass through 
the mesh into the codend (see Broadhurst and 
Kennelly, 1996). In the Hunter River (HR) prawn- 
trawl fishery (Fig. 1), the abundance of jellyfish 
means that commercial fishermen use blubber chutes 
throughout most of the trawling season. 
In a series of experiments that examined the per- 
formance of several types of BRD’s (Broadhurst et 
al., 1996; Broadhurst and Kennelly, 1996), we showed 
that a rigid separator-panel (the Nordmpre grid) sig- 
nificantly reduced the mean weight of bycatch in two 
estuaries and had no effect on the catches of prawns. 
Compared with the commercially used blubber chute, 
the Nordmpre grid also retained significantly less 
bycatch but caught more prawns. 
Bycatch-reducing devices, such as the Nordmpre 
grid and the blubber chute, function by mechanically 
partitioning the catch according to size (see 
Broadhurst et al., 1996), and therefore are generally 
not as effective in excluding unwanted individuals 
that are of a similar size or that are smaller than the 
targeted prawns. Previous studies have shown, how- 
ever, that it may be possible to exclude these smaller 
individuals by exploiting behavioral differences be- 
tween some species of fish and prawns (Watson et 
al., 1986; Broadhurst and Kennelly, 1994, 1995; 
Broadhurst et al., 1996). For example, studies by 
Watson et al. (1993) in the Gulf of Mexico showed 
that small individuals of red snapper ( Lutjanus 
campechanus), Atlantic croaker ( Micropogon 
undulatus), Atlantic bumper ( Chloroscombrus 
chrysurus ) and whiting (Menticirrhus sp.) were pas- 
sively excluded from trawls by various BRD designs 
comprising strategically placed panels of netting and 
escape exits. These designs were located posteriorly 
to a larger mechanical separating grid (designed to 
exclude turtles) and effectively functioned as second- 
ary BRD’s. 
It is apparent that several options exist for ways 
of excluding bycatch from prawn trawls. In the 
present study we wanted to determine which of these 
various devices (i.e. the Nordmpre grid, blubber 
chute, or some type of secondary BRD) is most ap- 
propriate for use in the HR prawn-trawl fishery. Our 
Figure 1 
The location of the Hunter River in New South Wales. 
specific goals, therefore, were 1) to assess the perfor- 
mance of four secondary BRD’s located behind the 
Nordmpre grid (including designs previously tested 
in the Gulf of Mexico by Watson et al., 1993) in re- 
ducing smaller unwanted individuals in the HR 
prawn-trawl fishery; 2) to compare the two most ap- 
propriate secondary BRD’s from 1) against a stan- 
dard Nordmpre grid and the commercially used blub- 
ber chute; and 3) to test a standard Nordmpre grid 
(with no secondary BRD) against the commercially 
used blubber chute. 
Materials and methods 
Two experiments were performed on commercial 
prawn-trawl grounds in the Hunter River (32°53'S, 
151°45'E, Fig. 1), between November and December 
1995 with a chartered commercial prawn-trawler 
(12.72 m). Three Florida flyers (mesh size=40 mm), 
each with a headline length of 9.14 m, were rigged 
in a standard triple gear configuration (see Andrew 
et al., 1991, for details) and towed at 2 knots across 
a combination of sandy and muddy bottoms in depths 
ranging from 2 to 8 m. Each of the identical outside 
nets were rigged with zippers to facilitate changing 
the codends (see Broadhurst et al., 1996). Because 
the middle net was not rigged in an identical man- 
ner to that used on the outside nets, its catch was 
excluded from analysis. 
The codends used in the experiments measured 50 
meshes long (2 m) and were constructed from 40- 
