480 
Fishery Bulletin 95(3), 1 997 
320 
M. undulatus 
240 
a> 
o 
c: 
CO 
"O 
c 
.Q 
CO 
£ 
Date 
Figure 1 B 
Nightly mean densities (three tows each night) of larval Atlantic croaker, 
Micropogonias undulatus , and pinfish, Lagodon rhomboides, caught near 
Beaufort Inlet, North Carolina. 
dentatus and Micropogonias undulatus , most species 
required a fourth-order polynomial, and Breuoortia 
tyrannus required a fifth-order polynomial. The poly- 
nomial models provided a means of estimating each 
species’ density for each day as well as the difference 
between the observed density and the estimated den- 
sity (residual ). One would expect that if an environ- 
mental factor influenced observed density, it would 
do so in a proportional sense, i.e. its effects would be 
exhibited in relation to the expected density of the 
species at that date during the season of recruitment 
for that species. We therefore divided each residual 
by the expected density for that date to obtain a 
measure that took the species’ recruitment pattern 
into account in looking at correlations with environ- 
mental variables. Durban-Watson statistics and first 
order autocorrelation coefficients were also computed 
to test the presence of autocorrelation and to mea- 
sure its magnitude. 
