302 
Fishery Bulletin 112(4) 
Table 1 
Animal and event details for 15 mortalities, for which complete data sets were available, detected in juvenile Steller sea 
lions ( Eumetopias jubatus) in the Gulf of Alaska between 2005 and 2013. Two values within a cell represent data returns 
from dual life history transmitters (LHX tags) implanted in sea lions, and, in the case of single values, only data from one 
tag was received. Mass at death was predicted from mean growth rates for age class and sex (Winship et al. 1 ) applied to 
mass at release. Postmortem temperature ( T ) is the mean of values recorded from time of death for 24 h or to onset of data 
transmissions from tags. In cases for which a temperature range for 24 h post mortem is not provided, tags began transmis- 
sions in less than 24 h. SST=sea-surface temperature. 
Animal no. 
52 
57 
51 
44 
54 
63 
35 
64 
46 
47 
58 
59 
32 
33 
62 
Month of death 
Aug 
Sep 
Oct 
Nov 
May 
Dec 
May 
Dec 
Jan 
Jan 
Feb 
Jan 
Feb 
Feb 
Mar 
Age at death (y) 
4.1 
2.3 
1.4 
2.4 
1.9 
1.5 
1.9 
1.5 
2.6 
1.6 
1.7 
2.6 
1.7 
1.7 
2.7 
Predicted mass at death (kg) 
243 
137 
110 
161 
96 
121 
163 
130 
171 
110 
116 
152 
158 
144 
148 
0.1 
0.6 
1 
1 
0.4 
5 
0.4 
0.1 
1 
0.2 
0.9 
0.6 
Days to transmit 
11 
0.1 
0.6 
1 
0.2 
1 
0.4 
9 
0.4 
0.1 
1 
0.2 
16 
0.9 
0.6 
10.4 
9.81 
8.58 
5.20 
5.07 
7.55 
4.96 
4.75 
3.96 
3.37 
3.27 
3.60 
T post mortem (°C) 
5.53 
10.5 
9.66 
8.41 
6.77 
7.02 
4.56 
7.54 
4.79 
4.54 
4.04 
3.42 
4.51 
3.16 
3.51 
2.2 
1.8 
0.0 
1.4 
T range 24 h post mortem (°C) 
0.0 
— 
— 
2.4 
— 
0.2 
— 
0.0 
— 
- 
0.8 
— 
Regional SST (°C) 
11.6 
10.6 
9.6 
8.2 
6.9 
5.1 
5.0 
5.0 
4.9 
4.7 
4.0 
3.7 
3.7 
3.4 
4.1 
SST source 2 
3,7 
4 
3,6 
4, 11 
3,6 
8, 13 
3,9 
8, 13 
3, 10 
3,5 
4,8 
7, 13 
12, 13 
13 
6, 13 
1 Winship, A. J., A. W. Trites, and D. G. Calkins. 2001. Growth in body size of the Steller sea lion (Eumetopias jubatus). J. Mam- 
mal. 82:500-519. 
2 Values from multiple sources were averaged (see Materials and methods section): 3 Comprehensive Large Array-data Steward- 
ship System; NOAA Data Buoy No. or identifier: 4 46060, 5 46061, 6 46076, 7 46080, 8 46081, 9 NKTA2, 10 SNDA2, and 11 VDZA2; 12 
GAK1 station of the Institute of Marine Science, University of Alaska Fairbanks; 13 first uplink from LHX tag. 
in Prince William Sound were averaged from values re- 
ported for locations in the Whale Bay, Eaglek Bay, and 
Zaikof Bay by Gay and Vaughan (2001). 
Results 
We detected 17 mortality events between 22 Novem- 
ber 2005 and 15 June 2013, over 40,620 cumulative 
exposure days in 36 implanted animals (28,202 days 
over the age range of 14-60 months). All 17 events 
were from animals with dual implants. Thirteen 
events yielded dual returns, and 4 events yielded sin- 
gle returns. Therefore, from the 34 tags implanted in 
the 17 animals that died, we recovered 27 full and 3 
partial data sets; 4 tags returned no data. We recov- 
ered 18 full data sets from 20 tags in 10 carcass sim- 
ulations; 2 tags yielded no data. Tags from carcasses 
uplinked on average within 32 days of deposition (SD 
14.2; range: 14-65 d; n- 18). From the ratio of dual 
(n= 21) to single (n=6) data returns from the 52 tags 
in 27 events, we estimated event detection probability 
at 98.4% (CI=95.0-100.0). This result indicates that 
we most likely did not miss any occurrences of mortal- 
ity within our sample of 36 animals implanted with 
transmitters. 
Of the 17 detected events, 2 yielded insufficient data 
to allow for inferences on causes of mortality. Complete 
temperature and diagnostic data sets from at least 1 
tag per animal were produced from each of 15 events 
(Table 1). Antemortem temperatures in all of thel5 de- 
tected mortality events with data were within normal 
ranges for healthy animals (36.0-38.2°C), and there as 
no evidence of pyrexia or hypothermia (Figs. 3-5). All 
of these 15 events occurred at sea and were classified 
as predation, as detailed later. The combined probabil- 
ity distribution of single to dual tag returns of 2+13 
for the 15 predation events and 2+8 for the 10 carcass 
tests that constitute simulated nontraumatic deaths 
gave an odds ratio of 1.625, indicating no differences 
in detection probabilities between predation events and 
simulated nontraumatic events (Fisher’s exact proba- 
bility test: P(2,d=1.0; Fisher, 1922). 
Of the 15 detected predation events, 11 events fell 
into a type-I scenario. In these events, temperatures 
dropped precipitously, with a mean peak AT of-0.81°C 
min -1 (SD 0.1), to ambient values that corresponded to 
estimates of regional SSTs (Fig. 3; Table 1). This rate of 
