Colura and Bumguardner: Effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry 
401 
All moribund specimens were classified as dead. Samples 
were transported to the Perry R. Bass Marine Fisheries 
Research Station (PRB) located near Palacios, Texas, 
where they were transferred to separate 60-cm diameter 
x 60-cm deep holding pens placed in an aerated 1500-L 
holding tank. Water in the holding tank was maintained 
within 2°C and 2 %o salinity of the bay temperature and 
salinity from which the samples were taken. Samples 
were checked at least hourly during transport to PRB. 
After having been placed in the holding tank, organisms 
were checked at least every three hours, except from 
2300 to 0700 hours, when they were checked once. Each 
time samples were checked, beginning when the samples 
were placed in transport tanks, all dead organisms were 
removed, labeled to identify sample, their time and date 
of collection were recorded, and the organisms were pre- 
served on ice for later examination. At approximately 1200 
hours on the day following collection of the sample (about 
21-27 h after collection), the remaining organisms were 
removed, their condition identified (living or dead), and 
the sample identification number and time and date of 
collection were recorded. All organisms collected in the 
sample were then identified, enumerated, and their total 
length (TL) was measured. 
The effect of salt-box use was evaluated for all bycatch 
species and for Atlantic croaker alone. Three measure- 
ments were used for the evaluation: percent survival of 
bycatch organisms when separation of catch was complet- 
ed (initial survival), percent survival of bycatch organisms 
at completion of sample observations (final survival), and 
mortality rate (M) of bycatch. Mortality rate was the ab- 
solute value of the slope of the regression of the natural 
log of living organisms +1 as a function of time. Time was 
measured from the time the catch was removed from the 
water to completion of sample observation. Because the 
three measurements had unequal variances, even when da- 
ta were transformed, nonparametric procedures were used 
to compare the two separation methods and to examine rela- 
tionships. Wilcoxon signed rank test was used to determine 
ifinitial survival, final survival, M, number of specimens col- 
lected, and TL of specimens differed significantly between 
the two separation methods. The Spearman correlation co- 
efficient and the probability that the coefficient was signif- 
icantly greater than 0 were used to examine relationships 
among all variables. For the correlation analysis, the use of 
a salt-box was coded as 1 and non-use as 0. 
Stepwise multiple regression analysis was used as an 
exploratory procedure to examine the effect of 1) length 
of time trawl was fished; 2) catch separation time (total 
time bycatch was onboard the vessel); 3) bay temperature 
at the time of collection; 4) bay salinity at time of collec- 
tion; and 5) effect of use or non-use of a salt-box on initial 
survival, final survival, and M of bycatch. For the regres- 
sion analysis, coding of the use and non-use of a salt-box 
was the same as that for the correlation analysis. The sig- 
nificance level of the independent variable to enter and 
stay in the model was P<0.05. Statistical analyses were 
performed by using the Statistical Analysis System (SAS 
Institute, 1990). All statistical tests were considered sig- 
nificant at the P=0.05 level. 
Effect of salt-box use on bycatch survival; control 
experiments 
Fourteen samples were collected from Matagorda Bay 
between 16 May and 25 October 1995, with a 6.1-m wide 
trawl constructed of 3.8-cm stretched nylon multi-filament 
mesh (Kana et al., 1993) equipped with a 6-nun stretch- 
mesh liner in the codend. The trawl was towed for 15 min- 
utes. The catch was removed from the trawl and placed 
on a 0.67 x 1.0 m table. The catch was allowed to sit on 
the table for seven minutes to simulate the time the catch 
sits on deck while commercial shrimpers re-initiate trawling 
operations. The 7-min period was selected on the basis of a 
preliminary survey that we conducted of Texas bay shrimp 
fishing practices. The catch was then divided into approxi- 
mately two equal portions: one half was placed in a salt- 
box and the bycatch was removed by skimming floating 
organisms from the surface; the remaining half of the catch 
remained on the table and shrimp were removed by hand, 
leaving the bycatch on the table. The salt-box was a 185-L 
fiberglass tank, half filled with bay water to which was 
added about 7.5 kg of food-grade salt. All bycatch removed 
from the salt-box was immediately placed in a separate 
89-L transport tank filled with bay water and supplied 
with compressed oxygen. In the hand separation process, 
shrimp were removed and the bycatch remaining on the 
table was placed in a separate 89-L transport tank filled 
with bay water and supplied with compressed oxygen. 
Dead organisms were removed from samples immediately 
after they were placed in transport tanks. Samples were 
then transported to the PRB where they were treated in 
the same manner as previously described for samples col- 
lected from commercial shrimpers. The effect of the salt- 
box on the bycatch (all species), and on Atlantic croaker, 
spot, and sand seatrout (each species separately) was eval- 
uated. Analyses were identical to those used to analyze 
samples collected from bay shrimpers. 
Results 
Bioassays 
Probit analysis indicated LE 50 for the five species ranged 
from a low of 17 min for red drum to 67 min for blue crab 
(Table 2). Southern flounder was the most tolerant of the 
fishes with an LE 50 of 58 min. Comparison of log-likeli- 
hood ratio G values indicated that NOEE ranged eight 
min for Atlantic croaker and red drum to 64 min for blue 
crab. Southern flounder had the greatest NOEE (32 min) 
of the four fishes tested. 
Effect of salt-box use on bycatch survival: samples 
from the bay shrimp industry 
All commercial fishermen resumed fishing operations 
before beginning the catch separation procedure. There- 
fore, for descriptive purposes, catch separation time can be 
divided into two periods: 1 ) time the catch remains on deck 
or table while fishing is resumed and 2) time the catch is 
