Stehlik et aL: Distribution of 3 predatory fish at a salinity front in a small estuary 
145 
Claypit Creek 
Depth, m 
was 2 
O Receiver 
—=1 Salinity front 
A, B Stations [ 
Sandy Hook Bay / 
40.40^N 
Northeast 
US 
x" ! 
Oceanic Bndge 
6 km 
McClees 
Creek 
40.38'"N 
Guyon Pt. 
Jones Pt. 
V/ New Jersey 
-) study area 
Swimming 
River 
Shrewsbury 
River 
40.36^N 
Navesink River 
Redk T1 km 
Bank Basin 
0 
2 km 
74.10°W 
74.05°W 
74.00“W 
Figure 1 
Map of the Navesink River, showing the location of the river on the northeastern coast of the United 
States, depths at mean low low water, locations of receivers used to track ultrasonically tagged blue- 
fish iPomatomus saltatrix), striped bass (Morone saxatilis), and weakfish (Cynoscion regalis) in 2007 
(indicated with circles), approximate location of the salinity transition front at flood tide (indicated by 
the black line) in 2007, and locations of station A at the front at flood tide and station B at a nearby 
channel, where gill net sampling was conducted in 2007 (indicated with the letter A or B). 
that cast of the SBE 25 where the front was found was 
designated as station A (depth; 1-2 m at low tide) for 
gill net sampling. The location of station A changed de¬ 
pending on the hydrodynamics. When the tide changed 
that day, station A was relocated. Station B for gill net 
sampling was located approximately 2 km downriver in 
a nearby channel (depth: 3-7 m at low tide) and was 
alv/ays in the same location (Fig. 1). 
Freshwater discharge records were obtained from 
U.S. Geological Survey streamflow station 01407500 in 
the Swimming River west of Red Bank, New Jersey 
(data available at website; Manderson et al., 2014). 
Fish collections and diets 
To investigate predator diets and prey distributions, we 
used targeted gill net sampling during the 12 weeks 
of hydrographic surveys. Three replicate nets were de¬ 
ployed at each of the stations A and B, at peak of flood 
tide and at peak of ebb tide, twice daily. They were 
anchored close to the river bottom, and were soaked 
for 2 h in the morning and again in the afternoon of 
that same day. The gear and soak times were chosen 
to be the same as those employed in 1998 and 1999 by 
Scharf et al. (2004). Gill nets were 45.7 m in length 
by 2.4 m depth, had 6 panels of equal length (7.6 m) 
and various mesh sizes (1.3-7.6 cm^). After fishing for 
2 h, gill nets were retrieved. Fish and macroinverte¬ 
brates captured in each net were sorted, counted, and 
measured. Striped bass and weakfish were measured in 
total length (TL), and measurements of bluefish were 
taken in fork length and converted to TL. Weakfish and 
bluefish were assigned to either age-0 or age-l-i- (age 
1 or older) cohorts on the basis of analysis of length 
frequencies in the earlier study (Scharf et al., 2004). 
Weakfish and bluefish <300 mm TL in spring and fall 
and <250 mm TL in summer were classified as age- 
0. Relative abundances of fish species from the front 
and channel stations were compared by using Mann- 
Whitney tests (F<0.05). 
Stomachs of the targeted predators were removed 
and preserved in 70% ethanol. Stomachs <5% full and 
those containing only unidentified matter were counted 
as empty. Fish and invertebrate prey were identified 
to the lowest possible taxonomic level, weighed wet 
(to the nearest 0.01 g), and lengths (in millimeters) 
of intact prey items were recorded. The most impor¬ 
tant prey taxa by percent weight of all predators were 
pooled into 10 categories, including a category for un¬ 
identified fish or other organisms. Cluster analysis was 
performed by the least squares method on percentages 
of prey taxa by predator, age class, and season. How¬ 
ever, too few striped bass were collected to conduct any 
diet analysis. 
Gill net collections were used to identify typical 
distributions of dominant prey taxa. We chose the size 
limit for predator-vulnerable fish as <150 mm TL, on 
the basis of lengths of prey in stomach contents from 
