408 
Fishery Bulletin 99(3) 
Table 12 
Independent variables which entered stepwise multiple regression models describing bycatch (all species), Atlantic croaker, spot, 
and sand seatrout short-term final percent survival, or mortality rate (M) (or both) resulting from their separation from shrimp 
with or without the aid of a salt-box during controlled experiments. Included are y-intercepts (±SE), coefficients (±SE) of the 
independent variables, partial R 2 of each variable in parenthesis, and model R 2 . A space indicates the variable did not enter the 
model. 
Independent variable 
Dependent variable 
v- intercept 
Separation time 
Bay salinity 
Bay temperature 
Model R 2 
Bycatch 
final survival 
216.59 ±58.069 
-2.71 ±0.88 (0.13) 
-4.68 ±2.08 (0.21) 
0.34 
M 
0.12 ±0.03 
0.09 ±0.03 (0.27) 
-0.002 ±0.001 (0.11) 
0.38 
Atlantic croaker 
M 
-5.36 ±203 
2.25 ±0.45 (0.75) 
0.21 ±0.08 (0.08) 
0.83 
Spot 
M 
Sand seatrout 
final survival 
M 
-1.44 ±1.41 
213.31 ±67.36 
-39.75 ±12.90 
0.18 ±0.07 (0.20) 
-7.08 (0.29) 
1.61 ±0.47(0.37) 
0.20 
0.29 
0.37 
for the dependent variable initial survival. Time required 
to separate catch accounted for 75% of the variation ob- 
served in Atlantic croaker M, whereas bay temperature at 
time of sample collection accounted for the remaining 8% 
variation explained by the model. Bay salinity at time of 
sample collection explained 20% of the variation observed 
in M of spot. Bay temperature at time of sample collection 
explained 37% and 29%, respectively, of the dependent 
variables M and initial survival of sand seatrout. No vari- 
able met the P<0.05 significance level for entry into models 
to explain initial survival of Atlantic croaker, spot, or sand 
seatrout, or final survival of Atlantic croaker or spot. 
Discussion 
Use of a salt-box to separate bycatch, as practiced by Texas 
shrimpers, had little or no effect on short-term survival 
of bycatch. Measurements of bycatch (all species) initial 
and final survival were not significantly different for the 
bycatch separation methods in samples collected from bay 
shrimpers and from TPW trawls. Significant differences 
between SB and NSB Atlantic croaker initial and final 
survival of samples collected from the commercial fishery 
were not corroborated by the results of the TPW control 
experiments. Exposure time to hypersaline conditions in 
a salt-box is short, generally less than two min, and the 
no-observed-effect exposure for selected species suggests 
that longer exposure (at least eight min for red drum and 
Atlantic croaker) would probably be required to signifi- 
cantly affect survival of most bycatch species. For these 
reasons, regulation of salt-box use as currently practiced 
by Texas bay shrimpers is unnecessary. 
The use of a salt-box had little impact on bycatch M, and 
initial or final survival. Salt-box use was correlated to M, for 
samples collected from the fishery, but not to initial or final 
survival. Salt-box use was not correlated to any of the three 
variables for bycatch in the control experiments. Atlantic 
croaker final survival was negatively correlated to the pres- 
ence of a saltbox in samples collected from the fishery but 
the correlation was not present in the control experiments. 
The salt-box variable entered the final survival regression 
models for combined catch and Atlantic croaker from the 
fishery but entered no other regression models. 
Environmental factors at the time of sample collection 
played a role in determining survival of bycatch, although 
not as great as other variables associated with bycatch 
separation. Both bay temperature and salinity at the time 
of sample collection affected survival and M variables in 
the control experiments as demonstrated by the correla- 
tion and regression analyses. Bay temperature also ex- 
plained a small portion of the variation (3%) associated 
with M in trials conducted on samples collected from the 
fishery. For individual species, bay temperature and salin- 
ity, either individually or combined, explained part of the 
variation associated with the regression models for M of 
Atlantic croaker, spot, and sand seatrout, and for final sur- 
vival of sand seatrout. The effect of temperature on M in 
samples taken from the fishery and survival and M in con- 
trol experiments suggest there may be a seasonal effect on 
survival of bycatch. Failure to observe a greater effect in 
trials with the fishery samples probably is due to the effect 
of other factors associated with the fishery, such as long 
trawl times and catch separation times that mask the ef- 
fect of environmental factors. 
Catch separation time was one of the most important 
variables associated with bycatch survival. It was corre- 
lated to M, initial survival, and final survival for samples 
collected from the fishery and to M and initial survival in 
the control experiment. The variable entered multiple re- 
