Buckel and Conover: Movements, feeding, and daily ration of Pomatomus saltatrix 
667 
from one haul of each of the 60- and 30-m seines in 
1993. Temperature (thermometer), salinity (refrac- 
tometer), and dissolved oxygen (modified Winkler’s 
method) were measured at each time point. Periodic 
functions were used to obtain quantitative values of 
tide and light levels for statistical analyses. Tides at 
the study site are semidiurnal and the tidal ampli- 
tude is ~1.5 m. The state of tide (T) for any given time 
point was calculated from the following equation: 
T = cos((( 27 t/ 12.42) x time of day) - 0 ), 
where 0 = the time of high tide in radians. 
A value for illumination (watts/cm 2 ) was calculated 
for each time point on each specific date (Kuo-Nan, 
1980). 
The effects of light, tide, salinity, dissolved oxy- 
gen, and temperature on CPUE of spring-spawned 
bluefish were examined with a forward step-wise 
multiple regression. Data from 1992 and 1993 were 
analyzed separately and log p (x+1) transformed to 
remove heterogeneity of variances. 
Diel collections — gill net and surface trawl 
A 90 x 2 m (4-cm-stretch, 2-cm-square) monofilament 
surface gill net was used to collect spring-spawned 
YOY bluefish from offshore shoal areas. One end of 
the gill net was anchored approximately 30 m east 
of the northernmost tip of Croton Point at a depth of 
-3 m (low tide) and set parallel to the north shore of 
Croton Point (see Fig. 1). The nearest beach-seine 
site was -200 m away in the cove just south of the 
gill-net set location. Gill-net collections were made 
concurrently with diel beach-seine collections on 20- 
21 July (approximate mid-set times were 1500, 2100, 
0600), 4-5 August (2100, 0130, 0600), and 18-19 
August (1500, 2100, 0130, 0600) in 1993. Soak times 
always lasted two hours. After net retrieval, blue- 
fish were removed and immediately frozen on dry 
ice. Fish from gill-net collections were not used in 
the calculation of daily ration. Relative abundances 
of bluefish were calculated as the number of fish 
caught per hour of soak time (CPUE). 
A surface trawl collection (8.2-m head-rope, 6.7-m 
foot-rope, 0.9-m-opening height, 2.5-cm-mesh net, 
0.6-cm-mesh codend) towed between two boats was 
made on 15-16 July 1993. It was conducted 1) to 
supplement night beach-seine collections, 2) to de- 
termine YOY bluefish movement patterns, and 3) to 
estimate daily ration. Collections were made every 
three hours beginning at 1230 and ending at 1230 
the following day. Two to three ten-minute tows were 
made at each time point. Tow speed was approxi- 
mately 4 knots. Bluefish and prey were immediately 
preserved in 10% formalin. Relative abundances of 
bluefish were calculated as the number of fish caught 
per trawl (CPUE). 
Diel collections — "movement collection" 
A combination gill-net and beach-seine collection was 
made on 11-12 August 1993 with the sole purpose of 
examining bluefish movement at crepuscular peri- 
ods. Fish from these collections were not used in the 
calculation of daily ration. Collections with beach 
seines were performed at 1200, 2400, and one hour 
before and after sunrise (0520 and 0720) and sunset 
(1900 and 2100). The temporal resolution of this sam- 
pling scheme with respect to sunrise and sunset was 
higher than that of evenly spaced intervals of beach 
seining described above (every three hours). Gill-net 
collections lasted two hours and were made through- 
out the diel cycle. 
Feeding period 
Values of gut fullness were used to examine the feed- 
ing periods of beach-seine-collected YOY bluefish. 
Gut-fullness values (F) were calculated as 
F = G/W, 
where G - prey wet weight; and 
W - bluefish wet weight (total weight minus 
prey wet weight; see “Diet analysis” 
below). 
Arc-sin square root and log t , (jc+ 1) transformations 
of individual gut-fullness values did not remove 
heteroscedasticity; therefore, the effect of time on gut- 
fullness values from beach-seine data (excluding 11- 
12 Aug and 11-12 Sep 1993) was examined with a 
nonparametric Kruskal-Wallis ANOVA. If treatment 
effects were significant, a nonparametric multiple 
comparison test for unequal sample sizes (Zar, 1984) 
was used to compare means. 
Daily ration estimates 
Values of gut fullness for spring-spawned bluefish 
from beach-seine (five dates in 1992 and four dates 
in 1993) and surface trawl (15-16 July 1993) collec- 
tions were used in estimating daily ration. Daily ra- 
tion was also estimated for summer-spawned blue- 
fish captured during beach-seine collections (19-20 
Sept. 1992). The Elliott and Persson ( 1978) food con- 
sumption model was used to estimate bluefish daily 
ration: 
