344 
Fishery Bulletin 109(4) 
cover. Unless indicated otherwise, all statistical tests 
were performed at the P=0.05 level of significance. 
Results 
Rates of at-vessel and postrelease mortality 
Capture date, sizes, deployment location, set pop-off 
date, ARGOS reporting location, days-at-liberty, and 
linear displacement for tagged sharks are summarized 
in Table 1. The overall PSAT reporting rate was 62% 
(CI*=50-73%), although reporting rates varied by spe- 
cies: 100% for silky sharks; 81% (CI* = 63-98%) for 
oceanic whitetip sharks; 50% (CI*=34- 65%) for blue 
sharks; 40% (CI*= 0-80%) for shortfin makos; and 38% 
(CI*=13-75%) for bigeye threshers. Median days-at- 
liberty were likewise species-specific: bigeye threshers, 
240 days (range: 181-240 days); shortfin makos, 165 
days (155-174 days); oceanic whitetip sharks, 164 days 
(10-243 days); blue sharks, 86 days (1-247 days); and 
silky sharks, 73 days (12-194 days). 
The fraction of sharks found dead during gear re- 
trieval was species-specific and was concordant with 
estimates derived from the commercial fishery (Table 2). 
More importantly, we were able to document only one 
case of postrelease mortality out of the 44 sharks (2.3%, 
CI*= 0-6.8%) whose PSAT transmitted data. One blue 
shark (male, 173 cm TL) expired seven days after being 
released (one mortality in 16 reporting tags affixed to 
blue shark; 6.3%, CI*=0-19%). 
Meta-analysis indicated the summary effect (Table 
3) for postrelease mortality in blue sharks was 15% 
(95% Cl, 8.5-25.1%). The Z statistic indicated that 
postrelease mortality was significantly different from 
zero (P<0.001) and the Q statistic indicated studies were 
measuring the same parameter (P= 0.680). Although the 
narrower 95% Cl bounds for the summary effect (Table 
3) indicated increased power over individual studies; a 
comparison of postrelease mortality estimates between 
Campana et al. (2009a) and the present study for blue 
sharks (with the assumption that sharks have equal 
chance of survival) at 80% power would require -275 
reporting PSATs (two-tailed Z-tests between two inde- 
pendent proportions at a=0.05, Zar, 1996). 
Horizontal movements 
For each of the pelagic shark species, estimated most 
probable tracks are shown in Figure 1. Error estimates 
for longitude were much lower than those for latitude 
in the movement model (Appendix 1). Geolocations could 
not be calculated for PSATs attached to bigeye threshers 
(Fig. IF) because of extreme and rapid vertical excur- 
sions coinciding at crepuscular times and the inability 
of the light sensor to record these changes (Musyl et 
al., 2001, 2003). 
For species where geolcation data were available, 
some individuals exhibited more directed movements 
as indicated by their advection-diffusion parameters 
whereas other movement patterns were more complex or 
cyclical (Fig. 1, Appendix 1). For example, the advection 
parameters for longitude (u=9.23) and latitude (v=3.84) 
indicated primarily east-west movements by the short- 
fin mako with ID 38572 (female, 210 cm TL) when it 
swam from subtropical waters near Hawaii to temper- 
ate waters in the North Pacific, including California 
Current coastal waters off central California (Fig. IE). 
Tagged silky sharks traveled west and southwest of 
the Hawaiian Islands in 20-2004, but south of 10°N in 
2005, near the North Equatorial Countercurrent (NEC) 
(Fig. 1C). The diffusion parameters estimated from the 
movement model indicated that six sharks exhibited 
relatively meandering swimming behaviors whereas 
three individuals (IDs 46585, 38581, 38573) exhibited 
more north-south directed movements. Oceanic whitetip 
sharks showed a complex movement pattern generally 
restricted to central Pacific tropical waters north of the 
NEC (Fig. ID). Nine individuals exhibited meandering 
swimming behavior, whereas three sharks (IDs 13113, 
38582, 46568) generally adopted more straight-line 
swimming modes, of which one shark (ID 38582) made 
a directed southward movement across the equator into 
the South Pacific. Although restricted to central Pacific 
tropical waters north of the NEC, blue sharks showed 
complex movement patterns (Fig. 1, A and B) from wa- 
ters near Hawaii into the Subtropical Convergence Zone. 
As determined from deployment and pop-up locations, 
blue sharks tagged in 2001 occupied latitudes from 
7.92° to 30.75°N (Fig. 1A), but individuals tagged in 
2002 did not travel farther south than 17.6°N (Fig. IB). 
Vertical movements 
Blue sharks remained significantly deeper and expe- 
rienced significantly cooler temperatures during the 
day than during the night (Fig. 2; Appendices 2 and 
3). Moreover, the significant daytime and nighttime 
differences in depth and temperature preferences were 
evident within and across individuals, and also when 
the data were grouped by sex (Figs. 2 and 3; Appendices 
2 and 3). As identified by the coefficient of variability, 
daytime and nighttime vertical movement patterns 
were similar, but the vertical movements of male blue 
sharks were significantly more variable than those of 
females. Coefficients of variability over 1.0 have been 
used to indicate possible mixtures in samples (Simpson 
et al., 1960) and values over 1.0 in blue sharks are 
reflective of individuals switching from a typical deep- 
daytime to shallow-nighttime vertical movement pat- 
tern, or exhibiting a mixture of the two patterns (Fig. 
2). The aggregated temperature-depth profile (Fig. 2B) 
shows that blue sharks regularly undertake movements 
beneath the uniformed temperature surface layer, but 
with considerable variability at crepuscular transitions 
(Fig. 2E). Several blue sharks adjusted their nighttime 
behaviors simultaneously with changing lunar illumina- 
tion (Appendix 2). 
Bigeye threshers showed the most striking differences 
in depth and temperature preferences and all MWBC 
