Table 1. Logistic Regression Results for Fish Tested with Brine Shrimp Nauplii

ERDC TN-DOER-E16

September 2004

RESULTS

Brine Shrimp Trials: The five fishes tested using brine shrimp exhibited different patterns in

their response to the different turbidity and prey concentrations. Results of the logistic regression

analyses are given in Table 1 and response surface information for each fish is presented in

Figures 1-5. Tests of the null hypotheses were rejected for all fishes (all p-values < 0.05). The

standardized estimates are useful for comparing the relative importance of the two factors

(turbidity and prey concentration) in influencing whether a fish had fed. For example, in the case

of pinfish, the two factors are nearly equivalent in the magnitude of their effects (standardized

estimates, Table 1), whereas in the case of spot, the effect of prey concentration was more than

twice the magnitude of the effect of turbidity.

Twenty percent of the menhaden fed during the trials (Table 1). Menhaden were most likely to

feed at low TSS and high prey concentrations (Figure 1). The probability that menhaden fed

dropped off sharply as both turbidity increased and prey concentrations decreased. More than

half (56 percent) of the pinfish fed, exhibiting a response similar in magnitude for the two factors

(Table 1). Some pinfish fed even at the highest TSS concentrations (Figure 2). Sixty-five percent

of fish spot fed during the experiment. At the highest prey concentration, there was little

reduction in the probability that fish had fed with increasing TSS concentration (Figure 3).

Turbidity had the strongest effect on feeding at the lowest prey concentration; however, the

effect of TSS was less than half that of prey concentration. Seventy percent of croakers fed, with

the TSS and prey factors of similar importance, but of lower magnitude than for the other fishes

(Table 1). As with spot, croaker had a high probability of feeding at the highest prey

concentration irrespective of the turbidity level (Figure 4). Thirty-nine percent of the flounders

fed, showing the strongest responses to both factors (Table 1) of any species tested. Flounders

were most likely to feed at high prey concentrations and low TSS concentrations, with a rapid

decline in feeding probability as these factors changed (Figure 5). Flounders did not feed at the

highest TSS by lowest prey concentration combination.

Table 1

Logistic Regression Results for Fish Tested with Brine Shrimp Nauplii

Species

Mean Fish Size (mm)

Fed

p-value TSS

p-value Prey

St. Est. TSS

St. Est. Prey

Menhaden

24.6

265

0.0001

-0.85

0.62

Pinfish

39.2

236

132

0.0001

-0.53

0.54

Spot

30.2

252

163

0.0003

0.0001

-0.32

0.70

Croaker

12.4

165

117

0.0129

0.0032

-0.25

0.30

Flounders

11.9

381

150

0.0001

-1.11

0.83

P-values indicate chi-square probabilities. Standardized estimates are given for both the total suspended sediments (TSS) and

prey factors.

Natural Plankton Assemblage Trials: The results of experiments in which plankton

assemblages were used as prey are not as easily interpreted because of lower replication and the

prey concentrations were a random rather than fixed factor in the statistical analyses. The

planktonic prey were comprised of several species and life history stages. Results of the logistic

regressions indicate that menhaden feeding success was significantly and inversely related to

TSS concentrations, whereas it was not significantly related to the prey assemblage