182 
Fishery Bulletin 1 14(2) 
PER tracks - Ebb tide 
I Beginning of track 
PER tracks - Flood tide 
I Beginning of track 
■I End of track 
X Location of predation event 
0 50 100 200 300 400 
H End of track 
X Location of predation event 
I Meters 
ejlI ■ 
Survival per zone - Ebb tide 
i : 0.99 - 1 
0.98 - 0.99 
0.97 - 0.98 
■HO 96 -0.97 
Total cumulative survival: 0.72 
Survival per zone - Flood tide 
CP 0.99 - 1 
r~ : 0.98 - 0.99 
M 0.97 - 0.98 
■i 0.96 - 0.97 
Total cumulative survival: 0.84 
Figure 2 
Multipanel aerial photograph of the study site. (A) Individual PER tracks during ebb tide condi- 
tions. Red x’s denote individual locations of predation. (B) Individual PER tracks during flood tide 
conditions. Red x’s denote individual locations of predation. (C) Study site segmented into 100-m 
sections. Color coding denotes survival per 100-m sections during ebb tide conditions. (D) Study 
site segmented into 100-m sections. Color-coding denotes survival per 100 m sections during flood- 
tide conditions. 
test of the null hypothesis that the P’s are zero was re- 
jected (LR=11.3, 2 df, P=0.004). The estimated distribu- 
tion of survival times was calcidated at the mean val- 
ues of the covariates (Fig. 3). These indicated that the 
proportion of salmon that were preyed upon increased 
sharply from 20 to 30 minutes of exposure to predators. 
We plotted the distribution of survival times as they 
varied from the minimum (negative) to the maximum 
(positive) water velocities by 0.1 m/s increments (Fig. 
4). Predation was greatest with increasing positive wa- 
ter velocities and was lowest at the more negative wa- 
ter velocities. 
From the PERs that were outfitted with cameras, 
we obtained 48 complete videos of individual deploy- 
ments (22% of total) that resulted in approximately 
800 gigabytes of raw data. Of these, tripped timers 
in combination with missing smolts indicated that 
12 were predation events. Video analysis confirmed 
the predation and a fish was seen preying upon the 
smolt in each instance. Three of the events captured 
on video were confirmed to be predation by striped 
bass (Morone saxatilis), and the remaining 8 predators 
were not identifiable to species. Analysis of the video 
data revealed that if the timer was activated and the 
