482 
Fishery Bulletin 95(3), 1997 
a 
>. 
Date 
Figure ID 
Nightly mean densities (three tows each night) of larval summer flounder, 
Paralichthys dentatus, and striped mullet, Mugil cephalus, caught near Beau- 
fort Inlet, North Carolina. 
before; in some cases the difference between two 
consecutive nights was an order of magnitude (e.g, 
Leiostomus xanthurus on 18-19 March). 
Although periods of low and high abundance can 
be seen along the time axis for each species, oscilla- 
tions in abundance by each species did not occur at 
the same time in all cases. However, most of the spe- 
cies share periods of high abundance (mid January, 
mid February, early March, and mid March). We pre- 
sumed that our data set would be a good candidate 
for time-series analysis. Spectral analyses of the es- 
timated densities for each species was employed to 
reveal evidence of periodicities of varying length. All 
eight species exhibited a strong 14-day signal, most 
likely dominated by the lunar cycle. However, this 
14-day signal turned out to be an artifact in the sam- 
pling method which was unavoidable. This artifact 
was due to sampling 50 min later each night in the 
sampling scheme (sampling at the same stage in the 
flood tide) until dawn, at which time the next sam- 
pling opportunity would be either 12 h 25 min later 
or 37 h 15 min later. We chose the later option (to 
