Lear et al.: Fine-scale behavior and habitat use of Pristispectinata 
353 
-a 
P a? — 
00:00 12:00 00:00 12:00 00:00 12:00 00:00 
May 10 May 11 May 12 May 13 
Date and time 
Figure 4 
Representation of acceleration and environmental data from an acceleration data logger 
that was attached to a young-of-the-year smalltooth sawfish (Pristis pectinata ) caught in 
the Peace River, Florida, and tagged on 9 May 2014 (sawfish ID 1, see Table 1). Vertical 
shading indicates periods of day (white) and night (light gray). Temperature and depth 
values show a diel pattern of depth use in which this individual remained in relatively 
shallow depths during the day and made repeated ventures to deeper water at night. 
Overall dynamic body acceleration (ODBA) was used to establish activity (dark gray 
bars) and burst patterns (black bars) that depict putative foraging activity; white bars 
indicate resting periods. 
short (5-10 s) bursts accompanied by sharp ascents 
(Fig. 3B), accounting for approximately 24% of observed 
burst events, and bursts of variable duration up to 45 s 
without predictable changes in depth (Fig. 3C). Water 
temperatures during deployments were 24.5-32.7°C for 
sawfish tagged in May and June and 20.6-28.9°C for indi¬ 
viduals tagged in November. 
The model that best described activity included depth, 
temperature, time of day, and tide as fixed-effect predic¬ 
tors (Table 2). Smalltooth sawfish were significantly more 
active during the evening and night hours, with a peak 
in activity around 2200, a few hours after typical sunset, 
and a second smaller peak around 0500, near sunrise 
(Fig. 5). Activity increased with temperature. Individuals 
were also more active at depths around 0.8 m (Fig. 5), with 
decreased activity in shallower and deeper areas, and 
were more active during the lower half of the tidal cycle 
(i.e., second ebb, low tide, and first flood) than during the 
higher half of the tidal cycle (i.e., second flood, high tide, 
and first ebb) (Tukey’s HSD: P<0.05; Fig. 5). Hourly mean 
depth of sawfish was significantly shallower at low tide 
(mean: 0.57 m [standard error (SE) 0.04]) than at high 
tide (mean: 0.75 m [SE 0.04]) (ANOVA: P<0.001, df=l, 
P=11.94), although there was no difference in mean depth 
between day and night (ANOVA: P>0.05, df=l, F=2.40). 
The best-fit model describing the probability of a burst 
event included only age class as a predictor, with bursts 
more likely to occur in YOY smalltooth sawfish than in 
individuals age 1 or older. However, time of day was the 
most important predictor of burst frequency and the only 
predictor maintained in the best-fit model (Table 2). This 
result indicates that, although the likelihood that bursts 
would occur was the same for all hours of the day, the 
burst frequency increased substantially during evening 
and night hours, with the highest burst frequency observed 
around 2200, coinciding with the peak in percentage of 
time sawfish were active. 
The distributions of the recorded depths of YOY and 
individuals age 1 and older were significantly different 
(Kolmogorov-Smirnov test: PcO.0001). An overall mean 
depth of 0.5 m was observed for YOY, with an hourly 
mean depth range of 0.1-1.5 m. Larger individuals (age 1 
or older) made occasional excursions into deeper water, 
with an overall mean depth of 0.8 m and an hourly mean 
depth range of 0.2-3.6 m (Fig. 6). 
Although acoustic tags were concurrently deployed with 
ADLs on all smalltooth sawfish, detections were sparse 
for some individuals during ADL deployments, with loca¬ 
tion categorized in 226 h of ADL deployments (126 h in 
the main stem of the Peace River and 100 h in mangrove 
creek habitats). This lack of detections for some fish made 
it difficult to link specific locations with behaviors, but 
data were sufficient to compare behaviors between broad 
habitat categories. Habitat type was an informative pre¬ 
dictor of activity, along with tidal period and depth, with 
individuals more active in the main channel than in 
mangrove creek habitats (Tukey’s HSD: P<0.05; Fig. 7). 
Habitat type was also an informative predictor of burst 
frequency, along with tide and depth, and of the burst prob¬ 
ability, along with tide, depth, and age class. Smalltooth 
