46 
Fishery Bulletin 99(1 ) 
Tagging operations certainly represent a stress for the 
fish, especially when performed after a traumatic capture 
and removal from the water to attach the tag. If we assume 
that a fish considers this particularly large floating object 
(the tracking vessel) to be a shelter against the stress or 
possible injury caused by tagging procedures, the asso- 
ciation with the vessel could then be interpreted as an 
antipredator behavior (shelter from predator hypothesis, 
Suyehiro, 1952, cited in Freon and Misund, 1999). In fact, 
Block et al. (1992) and Brill et al. (1993) did observe badly 
injured fish swimming within a few meters of the surface 
(one Indo-Pacific blue marlin and one striped marlin, 
respectively), which corresponds to the swimming pattern 
exhibited by our fish when following the moving tracking 
vessel. If an injury or significant stress occurred during the 
capture or tagging operations, one could expect to see post- 
tagging antipredator behavior. The time delay between 
release and the onset of association behavior differs from 
one fish to another and ranges up to 16 hours after release 
(tuna 1), which argues against a stress-related associa- 
tion caused by the tagging operation. Moreover, observa- 
tions of a school of yellowfin tuna exhibiting the same 
association as yellowfin tuna 3 prove that nontagged and 
apparently noninjured and nonstressed tuna develop the 
same association. Our observations thus do not support 
the “shelter from predator” hypothesis as an explanation 
for the attraction of tuna to the tracking vessel. 
The role of social behavior to explain the association 
of fish with floating objects has been expressed in the 
“meeting point” hypothesis (Dagorn, 1994; Dagorn and 
Freon, 1999; Freon and Misund, 1999). This hypothesis 
proposes the enhancement of fish aggregation by floating 
objects through improving the encounter rate between 
small schools or between isolated individuals, or both. Ac- 
cording to this hypothesis, tuna associate with various 
floating objects (drifting logs, anchored FADs, boats) to in- 
crease their chances of encountering conspecifics. Yellow- 
fin tuna 1 and 2 seemed to be isolated during the tracking, 
whereas yellowfin tuna 3 was a member of a school. Yel- 
lowfin tuna 1 broke its association with the tracking ves- 
sel, joining individuals (observed by the echo-sounder) lo- 
cated under a FAD. It is not possible to know if the fish left 
the tracking vessel because of the FAD or because of the 
conspecifics. This observation, however, appears to support 
the “meeting point” hypothesis: this tuna and those of the 
aggregation benefited from their respective associations to 
find more conspecifics. Yellowfin tuna 3 was visually ob- 
served to be with a school during nights when the school 
swam close to the boat, and acoustically observed when it 
was associated with FAD3. The school was estimated to be 
composed of 80 individuals while associated with FAD3. 
Our observations were not precise enough to determine if 
new individuals joined the school during the 4-day experi- 
ment, nor if the school broke its association with the boat 
to join a group already aggregated to FAD3, in a manner 
similar to that shown by yellowfin tuna 1. However, we be- 
lieve that the present observations do not reject the meet- 
ing point hypothesis. 
Yellowfin tuna 1 and 3 left the tracking vessel to stay 
close to anchored FADs. It is difficult to know, however, if 
they broke the vessel association to associate with FADs 
or to join conspecifics located close to the FADs, or both. 
Contact with yellowfin tuna 2 was lost owing to a heavy 
rain. We do not know if the fish voluntarily broke off the 
association with the vessel or if it simply lost contact with 
the vessel. For instance, if the fish used the sound of the 
vessel to stay close, it is possible that the sound of the ves- 
sel was masked by the rain. 
Because it is very important to know why tuna associate 
with floating objects (or vessels in the present case), it is al- 
so essential to understand why tuna leave floating objects. 
Although the sample size of our study was very small, it 
seems that the presence of other floating objects, conspecif- 
ics, or bad sea conditions can be responsible. Understand- 
ing the reasons why tuna form and break off aggregations 
is of major importance when studying the consequences of 
aggregation on tuna movements and distribution (Dagorn 
and Freon, 1999). 
Future studies 
The objective of a sonic tagging experiment is to observe 
movements of a fish in its natural environment. The vari- 
ety of experiments conducted throughout all the tropical 
oceans (Cayre and Chabanne, 1986; Holland et al., 1990; 
Cayre, 1991; Cayre and Marsac, 1993; Marsac et al. 1996; 
Bach et al., 1998; Josse et al., 1998; Marsac and Cayre, 
1998; Brill et al., 1999; Dagorn et al., 2000) have contrib- 
uted to a considerable increase in knowledge on the behav- 
ioral ecology of tropical tunas. Nevertheless, when a fish 
associates with a tracking vessel, although a very rare 
event, this objective has been violated. Among the track- 
ing experiments on 14 yellowfin tuna in French Polynesia, 
the distinction between vessel-associated and unassoci- 
ated individuals was very obvious. Moreover, this striking 
behavior has never been observed on other tuna species 
during tracking experiments (i.e. bigeye tuna, Thunnus 
obesus, and skipjack tuna, Katsuwonus pelamis). We con- 
sider that there is no possible doubt on the nature of 
the movements exhibited by a tagged individual (free 
movements or patterns associated with the vessel), which 
insures the validity of the interpretations of sonic tagging 
results. However, these rare events can be used to study 
the associations of fish with floating objects in a general 
sense. Rather than interrupting the tracking operation, 
we propose to develop particular experiments to improve 
our knowledge on tuna behavior. During the associations 
with vessels described in this paper, fish sometimes fol- 
lowed the vessel at speeds of up to 5 knots (2.6 m/s). It 
could be useful to use this behavior to study in situ the 
relationship between endurance time and velocity. More- 
over, it is important to collect data on the duration of asso- 
ciations. Observations of the biological environment of the 
associated fish would also be very useful to test the valid- 
ity of certain concepts, such as the “meeting point” hypoth- 
esis. During our experiments, we observed the biological 
environment (i.e. both prey and conspecifics) using an 
echo-sounder. The sounder assisted us to observe patches 
of prey (for yellowfin tuna 1) and the tuna aggregation 
