Musyl et al.: Postrelease survival, vertical and horizontal movements, and thermal habitats of five species of pelagic sharks 
355 
Table 4 
Cumulative percentage of temperature readings from pop-up satellite archival tags (PSATs) attached to silky and oceanic 
whitetip sharks expressed as differences from daily calculated sea surface temperature (4SST°C) for daytime and nighttime 
diving behavior. 
Silky shark (Carcharhinus falciformis) 
Day 
52.56 
87.06 
95.78 
98.26 
99.53 
99.84 
99.96 
100 
Night 
63.30 
91.63 
96.93 
98.80 
99.26 
99.48 
99.71 
99.83 
100 
Total 
57.71 
89.25 
96.33 
98.52 
99.40 
99.66 
99.84 
99.92 
100 
4SST (°C) 
0 
-1 
-2 
-3 
-4 
-5 
-6 
-7 
-8 
Oceanic whitetip shark ( Carcharhinus longimanus) 
Day 
63.89 
91.53 
96.88 
98.59 
99.36 
99.63 
99.87 
99.95 
100 
Night 
61.40 
89.98 
95.92 
98.39 
99.21 
99.58 
99.86 
99.94 
100 
Total 
62.67 
90.77 
96.41 
98.49 
99.29 
99.61 
99.86 
99.95 
100 
Methods for determining postrelease mortality 
in large pelagic fishes and sharks 
Implementing survival studies for pelagic species is chal- 
lenging because of logistics, cost, experimental design, 
and obtaining sufficient samples. There are only a few 
methods for estimating survival, and each has limita- 
tions. Historically, long-term survival of pelagic species 
has been estimated by large-scale conventional tagging 
programs with low return rates (<5%, blue shark, Kohler 
et al., 1998; ~1%, blue marlin, Ortiz et al., 2003). Such 
results are consistent with a high postrelease mortality 
but could also be attributed to large population sizes, 
dispersal, tag loss, or uncooperative fishermen. Direct 
observation in tank or pen studies (e.g., Mandelman and 
Farrington, 2007) may not be practical for large pelagic 
species. Although they are the right tool to indicate 
postrelease mortality, the cost of PSATs precludes their 
widespread application. Moyes et al. (2006) introduced 
a biochemical approach that reduces experimental bias 
and increases sample size and would therefore optimize 
experimental design. Once the method is operational, 
about 40 samples can be assayed for the cost of one 
PSAT (~US$ 4000) (Musyl et al., 2009). Other poten- 
tial methods that could achieve sufficient sample sizes 
have shown promise for other species (e.g., reflex action 
mortality predictors; Davis, 2007), but it is not known 
how well these methods would translate for large pelagic 
species. For example, we used the absence of movement 
in the nictitating membrane to determine at-vessel 
mortality, but it is not known whether variability in this 
response (or other responses) would be useful to predict 
postrelease mortality. Lastly, bioelectrical impedance 
analysis (Cox and Heintz, 2009) may be feasible if body 
condition correlates with long-term survival. 
Factors that influence mortality 
Presumably the effects of stress and injury during cap- 
ture are additive and unless there are overriding factors, 
we suggest that, under similar conditions (and with 
adequate sample sizes), the at-vessel and postrelease 
mortality rates for pelagic species should be roughly 
concordant (e.g., Moyes et al., 2006; Campana et al., 
2009a). For survival studies on blue sharks, the at-vessel 
and postrelease mortality estimates show close agree- 
ment. For example, Campana et al. (2009a) reported 
16% at-vessel and 19% postrelease mortality rates and 
we reported 5.9% at-vessel and 6.3% postrelease mor- 
tality rates. Although we did not find this relationship 
for other pelagic sharks, it is possible our sample sizes 
were not sufficient to detect differences between these 
two mortality rates. 
Clearly additional research is required to determine 
whether at-vessel mortality correlates with postrelease 
mortality across a range of shark species and to deter- 
mine which biological and anthropogenic factors account 
for variability in mortality estimates. As discussed in 
Musyl et al. (2009), postrelease mortality estimates in 
Campana et al. (2009a) may have been strongly influ- 
enced by handling. Hoey and Moore 1 suggested that a 
20% difference in mortality for blue sharks discarded 
from longlines was attributable to handling practices in 
the Atlantic fishery where the Campana et al. (2009a) 
study took place. Campana et al. (2009a) also reported 
a significant vessel effect in their survival model of 
retrieved dead sharks, which the authors attributed to 
handling. Carruthers et al. (2009) and Diaz and Serafy 
(2005) also suggested discard and handling practices 
may have been responsible for differences in at-vessel 
mortality rates of blue sharks in the Atlantic longline 
fishery. If differences in handling practices strongly cor- 
relate with variable survival, a logical extension would 
be to develop discard-and-release regulations that could 
significantly improve survival (Carruthers et al., 2009). 
Circle hooks were used throughout our study which 
probably increased both the at-vessel survival (Diaz and 
Serafy, 2005; Kerstetter and Graves, 2006; Campana 
et al., 2009a; Carruthers et al., 2009; Musyl et al., 
2009) and postrelease survival of blue sharks (Moyes 
