Fabrizio et al. : FHome range and seasonal movements of Centropnstis striata 
85 
Transmitters (V8SC-2H, Vemco, Ltd., Bedford, Nova 
Scotia, Canada) were implanted in 129 anesthetized 
Black Sea Bass (mean size, 307 mm TL; range, 220- 
431 mm TL; 72 = 129 ) at the study site using surgical 
techniques described in Fabrizio and Pessutti (2007). 
Dummy transmitters of the same size (30 mm long and 
9 mm in diameter), shape, and weight (5 g in air, 3.1 g 
in water) had 100% retention rates in laboratory-held 
Black Sea Bass (Fabrizio and Pessutti, 2007). Trans- 
mitters emitted a coded acoustic signal at 69 kHz ev- 
ery 210 s on average (signal delay varied randomly be- 
tween 120 and 300 s) and had a battery life of 384 d. 
Surgery time averaged 4.1 min (range, 2.0-11.7 min; 
n=128 [surgery time not recorded for 1 fish]); immedi- 
ately after surgery, the length and sex of each fish was 
recorded. Sex determination was limited to classifica- 
tion of fish as either male (n=34) or fish of unknown 
sex (n=88); the latter group included females, subordi- 
nate males, and transitional males. 
Sizes of the 2 groups overlapped: males ranged be- 
tween 270 and 431 mm TL, with a mean of 343.5 mm 
TL (+standard error of the mean [SE] 6.32) and fish 
of unknown sex ranged between 220 and 395 mm TL, 
with a mean of 292.6 mm TL (SE 4.50). We used ram 
ventilation to resuscitate fish in a flow-through seawa- 
ter tank onboard the vessel and released fish within 
the study site as soon as fish were able to swim force- 
fully downward. All procedures were conducted at a 
mean temperature of 17.8°C (range: 13.6-24.2°C) and 
27.1 psu salinity (range: 22.1-31.0 psu). Of the 129 im- 
planted fish, 5 carried transmitters that malfunctioned 
and 2 fish died within hours of release, resulting in 
122 live Black Sea Bass with functioning transmitters. 
detected within the study site, and this estimate is 
likely a conservative one. In particular, fish captured, 
tagged, and released at the end of the tagging period 
may have previously occupied the site. 
Release date was considered in the model because 
preliminary investigations revealed potentially sig- 
nificant variation in home-range size associated with 
this factor. The model, which contained only fixed ef- 
fects, was fitted using the MIXED procedure in SAS 
(vers. 9.3, SAS Institute, Inc., Cary, NC), and model 
parameters were estimated with restricted maximum 
likelihood (REML) (Littell et al., 2002). Examination of 
residuals from an initial model run indicated heteroge- 
neity of variance in the home-range data; we, therefore, 
log e -transformed home ranges and verified that residu- 
als from this model supported the assumption of homo- 
geneity of variance. To test the assumption of equality 
of slopes for males and fish of unknown sex, we fitted 
a model that included all possible 2-way interactions 
with sex (sexxduration of occupancy, sexxlength, and 
sexxrelease date): 
Y iik =p + S + y + a i + p j + (a8) i + ( ay) i + (a/J)^ + e ijk , 
where Y;j k 
A 
5 
Y 
Pi 
f ijk 
the natural log of the home range for the 
k th fish of the i th sex released on the j th 
date; 
the intercept; 
the effect of duration of occupancy (days); 
the effect of fish size; 
the effect of the i th sex; 
the effect of the j fch release date; and 
the random unexplained error. 
Home range 
Indices of movement 
Home-range area, measured in hectares (ha), was cal- 
culated for individual Black Sea Bass with the kernel 
density estimator. Kernel methods provide a probabi- 
listic description of the space used by an organism and 
require a smoothing parameter to efficiently describe 
the density distribution of the individual location data 
(Worton, 1989). We used the bivariate normal distribu- 
tion to estimate kernel home ranges with the adehabi- 
tat package in R (R Development Core Team, 2005; Ca- 
lenge, 2006). For each fish, we eliminated observations 
from the release date to exclude data from the period 
during which fish may have been recovering from sur- 
gery and excluded data from the last week of occupan- 
cy at the site to ensure that we considered only those 
fish that were clearly not in the process of dispersing 
from the study site (Fabrizio et al., 2013); as a result, 
109 fish were retained for analysis of home range. 
We used the 85% probability polygon to define home- 
range area and investigated the effect of sex on home 
range through the use of an analysis of covariance with 
fish size, duration of occupancy (days), and release date 
as covariates. Duration of occupancy was estimated on 
the basis of the total number of days that the fish was 
To gauge movement of fish, we developed 2 indices. The 
binomial movement index allowed us to investigate 
factors associated with movement (e.g., are fish more 
likely to move during the night?), whereas the continu- 
ous activity index provided information on the extent 
of movements (e.g., how does activity level change in 
response to temperature?). The movement index indi- 
cated whether fish moved between adjacent acoustic 
stations during a given 3-h interval on a given day; 
this index was coded “0” (no movement) or “1.” 
To minimize autocorrelations among hourly observa- 
tions, we considered location data during 4 time peri- 
ods of each day; dawn, day, dusk, and night. Dawn and 
dusk periods were defined as the 3 hours centered on 
the beginning and end of nautical twilight (U.S. Naval 
Observatory, http://aa.usno.navy.mil/faq/docs/RST_defs. 
php); day and night were defined as the 3-h periods 
midway between dawn and dusk or dusk and dawn. 
The movement index for each time period was calcu- 
lated daily for each fish between the time of tagging 
and 19 November 2003; after this date, fewer than 3 
fish were detected in a given time period. In this man- 
ner, 24,789 observations from 121 fish were available 
