Dagorn et al.: Association of Thunnus albacares with tracking vessels 
45 
seems, however, that this behavior is not size dependent 
(sizes ranged from 60 to 167 cm FL) nor is it related to 
the size of the tracking vessel: 12-m vessel for Cayre et al. 
(1996), 20-m and 53-m vessels for Brill et al. (1999), and 
28-m vessel in our study. 
Horizontal movements 
The horizontal movements of fish associated with tracking 
vessels duplicated the horizontal movements of the track- 
ing vessel. Therefore, the observed paths are not compa- 
rable with horizontal movements of tagged fish that were 
not associated with tracking vessels. Our results, however, 
give information on the duration of associations and pos- 
sible “competition” between FADs and tracking vessel to 
attract the tagged fish. 
Considering the different patterns of movements of tu- 
nas observed at anchored FADs, Holland (1996) proposed 
three horizontal patterns: 1 ) fish that leave the FAD and 
show no tendency to return to it over the duration of the 
track; 2) fish that spend the entire duration of the track 
(day and night) within a few hundred meters of the FAD, 
and 3) fish that spend daylight hours at the FAD site, 
leave at night and return to the same or an adjacent FAD 
the next day. Fish 1 and 2, as well as the 167-cm yellow- 
fin tuna tracked by Brill et al. (1999) near the main Ha- 
waiian Islands, associated with the tracking vessel during 
daytime, which corresponds to the third class defined by 
Holland (1996). Conversely, Cayre et al. (1996) reported 
a nighttime association between a 108-cm yellowfin tuna 
and the tracking vessel. Moreover, fish 3 remained associ- 
ated with the tracking vessel for more than two 24 h (con- 
tinuous day and night cycles) which corresponds to the 
second pattern defined by Holland (1996). Although our 
sample size was small, the three fish of our study, and 
the two other yellowfin tuna that exhibited such associa- 
tion (Cayre et al., 1996; Brill et al., 1999), exhibited dif- 
ferent lengths of associations, at different periods of the 
diurnal cycle, all of which also correspond to the variety of 
patterns observed for fish associated with anchored FADs. 
These features, however, cannot be used to determine if 
tuna treat drifting and anchored floating objects different- 
ly, as first proposed by Holland et al. ( 1990). 
The possible competition between FADs and the track- 
ing vessel to aggregate tuna is an interesting feature of 
our results. The three tuna were caught close to and were 
considered associated with a FAD. Yellowfin tuna 1 and 
2 left their FADs after release and did not associate im- 
mediately with the tracking vessel, whereas yellowfin tu- 
na 3 associated with the tracking vessel after release. In 
the last part of their track, yellowfin tuna 1 and 3 clearly 
abandoned the vessel to associate with FADs. The pres- 
ence of yellowfin tuna 2 was also noticed close to the FAD 
a few hours after the end of the tracking. Attempts to re- 
aggregate yellowfin tuna 3 while associated with the third 
visited FAD were not successful. However, it is noteworthy 
that this tuna chose to associate with the tracking vessel 
rather than to FADs 1 or 2 on the previous days, showing 
a different motivation than those exhibited toward FAD3. 
Cayre et al. (1996) attempted to abandon the associated 
108-cm yellowfin tuna by rapid vessel accelerations, trying 
to make the fish associate with a FAD, but without success. 
We cannot determine the reason for a possible preference 
of fish toward vessels or FADs. We can only propose that 
tuna regard the vessel and FADs in a similar manner, or 
that the choice in aggregating between these two struc- 
tures depend on factors (external or internal stimuli) that 
we could not record during our experiments. 
Vertical movements 
It is known that the swimming depth of yellowfin tuna is 
controlled by the diurnal cycle: surface swimming at night 
and deep swimming at daytime (Carey and Olson, 1982; 
Cayre and Chabanne, 1986; Holland et al., 1990; Cayre, 
1991). However, besides this diurnal behavior, it seems 
that fish travel closer to the surface when associated with 
a moving vessel: yellowfin tuna 1 and 2 swam very close 
to the surface (mean swimming depth=5.3 m ±2.9 for tuna 
1 and 8.4 m ±4.5 for tuna 2) when they were associated 
with the moving vessel during daytime. Yellowfin tuna 3 
also exhibited very shallow swimming depths when associ- 
ated with the tracking vessel, but during nighttime (mean 
swimming depth=11.3 ±6.6 and 9.3 ±4.8 for the first two 
nights). Cayre et al. (1996) and Brill et al. (1999) did not 
report any relationship between the swimming depth of 
the fish and the speed of the tracking vessel. Figure 4 indi- 
cates definite relationships between fish swimming depths 
and the speed of the vessel during the associations. How- 
ever, while this vertical reaction of the associated fish to 
the different vessel speeds have been observed during day- 
time, we should mention that no observations were made 
to examine the response of the fish to low vessel speeds 
during nighttime, which should have been deeper than 
the depths exhibited by yellowfin tuna 3 during the first 
two nights. This association behavior is similar to that 
observed by Holland et al. (1990), in that fish tend to 
be closer to the surface when associated with FADs. We 
propose that floating objects generally induce the fish to 
swim closer to the surface and that this tendency increases 
when floating objects are moving fast. 
In addition to the shallower swimming of yellowfin tuna 
when associated with moving vessels, the amplitude of ver- 
tical oscillations are drastically reduced. We suggest that 
when fish are associated with a vessel, they reduce the am- 
plitudes of their vertical oscillations, and that the mean 
swimming depth is partly controlled by the speed of the 
vessel (i.e. the distance from the fish to the tracking vessel 
decreases when the vessel speed increases). The reasons for 
this change are not known. More data are clearly needed 
to examine the exact effects of a floating object (including 
its speed) on the vertical pattern of associated fish and 
to distinguish these effects from those due to the diurnal 
cycle, thermoregulation, or foraging behavior. 
Hypotheses to explain why tunas associate or 
disassociate with tracking vessels 
The reasons why pelagic fish associate with floating objects 
are still not known conclusively (Freon and Misund, 1999). 
