194 
Fishery Bulletin 111(2) 
a = the week effect (z=l, 2, 3, ...32); 
5 = the time-of-day effect (/=day, night); 
t = the tidal stage effect (&=ebb, flood, slack be- 
fore ebb, and slack before flood); and 
y = the effect of mean temperature. 
All effects in this model were considered fixed. All 
plausible interactions (temperaturextidal stage, 
temperaturextime of day) were investigated and found 
to be insignificant (a=0.05). We modeled the repeated 
measures of activity (discrete count data) as a nega- 
tive binomial response after verification of the superior 
fit of the negative binomial distribution to the Poisson 
distribution to these data. According to the quasi-like- 
lihood information criterion (a modification of Akaike’s 
information criterion applied to models fitted by gen- 
eralized estimating equations; Littell et ah, 2006), the 
independent correlation matrix best described the na- 
ture of the correlation among repeated measurements 
within subjects. This correlation matrix is the simplest 
covariance model, where the within-subjects correla- 
tion is zero (Littell et ah, 2006). 
6 / 1/2007 8 / 1/2007 10 / 1/2007 12 / 1/2007 2 / 1/2008 4 / 1/2008 
Date 
Figure 2 
The probability that tagged Summer Flounder ( Paralichthys 
dentatus) resided in the Wachapreague lagoon system from 
June 2007 to April 2008 on basis of the Kaplan-Meier estima- 
tor (solid lines) with 95% confidence intervals (dashed lines). 
Dispersal of tagged fish (♦) was monitored from 8 June 2007 
until the last fish emigrated on 17 January 2008. Dispersal 
rates changed significantly after 11 October 2007 (change- 
point). The time before the change-point was considered the 
residency period, a time during which most fish remained 
within the lagoon system. The time after the change-point 
was considered the emigration period, during which most 
fish were observed dispersing from the Wachapreague lagoon 
system. Returns of tagged fish (■) were monitored from 18 
January 2008 to 29 July 2008. The last return was detected 
on 7 April 2008. 
Results 
Migratory behaviors 
The fish included in all subsequent analyses were 
the fish that were alive and detected at receivers as 
of 8 June 2007. As a result, 45 out of 50 tagged fish 
were included in the analyses (278-558 mm TL). Of 
the 5 fish we eliminated, 1 was assumed dead (all de- 
tections were at a single receiver). Another fish de- 
parted through the inlet before 8 June 2007 and was 
subsequently detected in Delaware Bay (-100 km to 
the north) on 9 June 2007 (Fox 2 ). The remaining 3 fish 
were never detected by our receivers, and we assume 
these fish either departed undetected or were harvest- 
ed by recreational anglers but not reported. All except 
1 of the 45 fish included in the analyses were detect- 
ed in June; the remaining fish was detected for the 
first time in July. Most tagged individuals accounted 
for <6% of the total number of detections, which was 
165,003. Two fish, however, contributed 24% and 13% 
of the total detections. These individuals were detected 
continuously at receivers for long periods of time with 
few gaps between detections. 
The mean residence time for Summer Flounder 
in the lagoon system was 130 ±13 days, or about 
4.3 months (range: 18-223 days). Fish dispersed 
throughout the study period, but dispersal rates 
increased significantly after mid-October [change- 
point=week 18 (11 October 2007); F=212.2, 
P<0.05; Fig. 2]. On the basis of the Kaplan-Meier 
estimator, only 27% of tagged Summer Flounder 
had dispersed by 10 October 2007. Accordingly, 
the period from 8 June 2007 to 10 October 2007 
was identified as the residency period (Fig. 2); the 
majority of fish that dispersed during this peri- 
od did so shortly after they were tagged in June 
(Fig. 3). June was also the month with the high- 
est number of censored fish (i.e., fish of unknown 
fate; Fig. 3). The emigration period was identified 
from 11 October 2007 to 17 January 2008, when 
the last fish departed (Fig. 2). During this period, 
dispersal rates increased such that 50% of fish 
departed by 11 November 2007, although most 
fish (31%) dispersed in December. Only 7 fish 
were classified as temporary emigrants, remain- 
ing undetected for 14 or more consecutive days 
(range: 14-154 days) after detection near the in- 
let and before redetection in the system and sub- 
sequent final dispersal. 
Between 7 February and 7 April 2008, 17 Sum- 
mer Flounder (36%) returned to the lagoon sys- 
tem (Fig. 2). Four of these fish did not disperse 
through Wachapreague Inlet in 2007, and, there- 
fore, their dispersal dates were unknown. Of the 
returning fish with known dispersal dates, 58% 
(11 individuals) dispersed during the emigration 
2 Fox, D. 2007. Personal commun. Delaware State 
Univ., 1200 N. DuPont Highway, Dover, DE 19901. 
