200 
Fishery Bulletin 96(2), 1998 
<n 
JO 
E 
o 
Z 
120 
100 
80 
60 
40 
20 
mean: 60.61 
median: 15 
n = 257 
I.. 
Liiu 
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 
Number of bycatch 
Figure 3 
Distribution of bycatch sampled during the 1995 gulf menhaden fishing season. 
Table 2 
Summary statistics of loglinear models examined through the stepwise selection procedure. Models with fits as good as the 
saturated model are marked ns. (A=area, B=bycatch, D=dolphins, S=season). Models are presented hierarchically from most 
complex to simplest. 
Loglinear model 
df 
G 2 
/-'-value 
Model without the four-way interaction term 
ADB SDB SAB SAD 
6 
11.64 
0.0706 
Selected models with 3 three-way interaction terms 
ADB SDB SAB 
12 
26.58 
0.0089 
ADB SDB SAD 
12 
27.62 
0.0063 
SDB SAB SAD ns 
9 
13.41 
0.1449 
SAB SAD ADB 
14 
27.90 
0.0147 
Selected models with 2 three-way and 1 two-way interaction terms 
SDB SAB DA 
15 
29.38 
0.0144 
SAB SAD DB ns 
11 
15.65 
0.1547 
Selected models with 2 three-way interaction terms 
SAB SAD ns 
12 
17.82 
0.1213 
SDB SAB 
18 
45.83 
0.0003 
Selected models with 1 three-way and 2 two-way interaction terms 
SAB SD AD 
18 
34.89 
0.0097 
SAD SB AB 
18 
34.30 
0.0116 
