Heftier et at: Variability in abundance of larval fishes inside Beaufort Inlet 
483 
Table 1 
Average seasonal density (=AveDen) and maximum daily density (=MaxDen) of target species listed in order of decreasing aver- 
age density (larvae per 100 m 3 ). Sampling was conducted between 20 November 1991 and 15 April 1992. 
Scientific name 
Common name 
AveDen 
MaxDen 
Capture date 
Leiostomus xanthurus 
spot 
82.8 
504.0 
21 Dec-15 Apr 
Micropogonias undulatus 
Atlantic croaker 
37.9 
274.2 
20 Nov-15 Apr 
Lagodon rhomboides 
pinfish 
30.0 
342.0 
22 Nov-15 Apr 
Brevoortia tyrannus 
Atlantic menhaden 
10.0 
159.8 
22 Nov-15 Apr 
Paralichthys lethostigma 
southern flounder 
3.6 
28.2 
04 Dec-11 Apr 
P. albigutta 
gulf flounder 
2.7 
19.7 
20 Nov-15 Apr 
P. dentatus 
summer flounder 
2.0 
12.2 
31 Dec- 15 Apr 
Mugil cephalus 
striped mullet 
1.0 
15.2 
04 Dec- 14 Apr 
Table 2 
Standard error of the mean abundance of larval species obtained at 2- to 30-day subsampling intervals of the actual daily sam- 
pling data set. 
Scientific name 
If the number of days between sampling had been 
2 
3 
4 
5 
7 
14 
30 
B. tyrannus 
0.78 
0.89 
0.93 
1.32 
1.98 
1.93 
1.54 
L. rhomboides 
2.60 
2.36 
1.62 
3.05 
4.14 
4.14 
3.89 
L. xanthurus 
0.73 
5.22 
2.52 
7.99 
5.68 
12.86 
9.50 
M. undulatus 
1.96 
2.28 
1.17 
1.37 
1.90 
5.90 
3.65 
M. cephalus 
0.31 
0.26 
0.22 
0.21 
0.38 
0.24 
0.33 
P. albigutta 
0.37 
0.14 
0.44 
0.19 
0.27 
0.26 
0.31 
P. dentatus 
0.07 
0.26 
0.14 
0.09 
0.19 
0.32 
0.38 
P. lethostigma 
0.15 
0.24 
0.43 
0.22 
0.42 
0.48 
0.46 
avoid sampling twice on the same date), but either 
option would have resulted in an irregular pulse in- 
terval in the time line every 14 days and would have 
precluded further time-series analysis. 
Within-night variability 
Variability between tows of the three tows each night 
was estimated by calculating the coefficient of varia- 
tion (CV) for each night and for each species. The 
CV averaged about 50% for each species throughout 
the sampling period. Late in the immigration period, 
the CV in densities within a night increased for 
Paralichthys lethostigma', for all other species daily 
tow-to-tow variability was relatively constant 
throughout the time series. 
Sampling interval 
The range in density estimates derived from sub- 
sampling the actual data set increased as the sub- 
sampling interval increased (Fig. 2), as did the stan- 
dard error of the mean for most species (except Mugil 
cephalus and Paralichthys albigutta) (Table 2). For 
example, sampling every night yielded a seasonal 
mean of 10 B. tyrannus larvae per 100 m 3 , but if we 
had sampled only every 30 days our estimate for the 
1991-92 immigration season, seasonal mean abun- 
dance could have ranged from 3 to 43 larvae per 100 
m 3 , depending on which date we began sampling. 
Similarly, had we sampled for L. xanthurus every 30 
days, our estimated seasonal mean could have ranged 
from 11 to 249 larvae per 100 m 3 . 
Daily versus weekly sampling 
Sampling methods were compared to see if the in- 
creased effort required for daily sampling provided 
better abundance estimates than did weekly sam- 
pling. Weekly estimates were calculated by using B. 
tyrannus abundance data from our daily 1-m-net 
oblique-tows and are plotted (Fig. 3) along with 
