Holland et al Movements of Thunnus albacares and T obesus near fish-aggregating devices 



505 



1985). Alternatively, it is possible that nighttime feed- 

 ing excursions away from the FADs, combined with 

 the apparently reduced daytime swimming activity at 

 FADs as compared with reef banks, result in an ener- 

 getics balance equivalent to that of better-fed, but more 

 active, open-ocean fish or reef-perimeter fish. Even 

 though feeding may not occur as often at FADs as in 

 other areas, FAD-associated tuna may conserve energy 

 normally associated with active hunting. By orienting 

 to FADs, the fish maintain a fixed position in the cur- 

 rent flow and may feed on current-borne prey, a situa- 

 tion analogous to trout holding station in a river. The 

 predominant upcurrent orientation of FAD-associated 

 tunas may serve to intercept incoming prey items 

 before they can use the FAD structure for shelter. 

 These feeding advantages would not pertain to tuna 

 associated with logs, which drift with the current and 

 cannot provide the function of serving as a fixed point 

 in a stream. Thus, although logs may indicate areas of 

 enriched forage for planktivores by marking zones of 

 Langmuir cell convergence (Fedoryako 1982), they are 

 unlikely to supply the high food demands of even a 

 medium-sized school of tuna, and, in fact, small fishes 

 found in association with logs are not an important part 

 of the diets of tuna taken from beneath logs (Hunter 

 and Mitchell 1967). 



There are other indications that different phenomena 

 may imderlie the association of tuna with natural debris 

 and their association with FADs. For instance, fishing 

 techniques which take advantage of log-associated 

 tuna indicate that in these circumstances the tuna tend 

 to aggregate around logs at night and move away in 

 the daytime— exactly the opposite of the behavior 

 observed in the current study. However, in the log 

 situations, tuna behavior may be perturbed by the use 

 of lights placed on the logs by fishermen. Also, since 

 tunas probably are capable of magneticaly mediated 

 orientation (Walker 1984), and in our own study ex- 

 hibited precise feats of navigation (e.g., returning to 

 the same FAD; navigating between FADs; maintain- 

 ing straight azimuths over prolonged periods of time), 

 it is unlikely that tunas have trouble distinguishing 

 between logs and FADs because, unlike logs, FADs are 

 both stationary and attached to the ocean floor by 

 visibly (and possibly, acoustically) discernible chain and 

 rope. 



The combined data yield an average nightime excur- 

 sion distance of approximately 5 nmi. This would in- 

 dicate that to maximize the ability of FADs to aggre- 

 gate yellowfin tuna consistently, and to minimize the 

 influence of FADs on coastline-associated populations, 

 FADs should be placed a minimum of 5 nmi from the 

 nearest 40-fathom bank. This interpretation is sup- 

 ported by the track of fish YF8504 which rapidly 

 traversed the 3.5 nmi to the 40-fathom contour when 



disturbed at R FAD by a school of porpoises. Similar- 

 ly, YF8405 was caught at R FAD 2 weeks after being 

 tracked along the adjacent coastline which, at its near- 

 est point, is between 3.5 and 4.0 nmi away. In fact, op- 

 timal FAD placement would be greater than 10 nmi 

 offshore (or from the nearest 40-fathom bank) so that 

 the radii of influence of the FAD and the coastline do 

 not overlap. Frequent anecdotal observations and the 

 impressions of Hawaiian fishermen suggest that FADs 

 located 10 nmi offshore are more consistently produc- 

 tive than FADs located closer to shore. 



Daytime orientation of yellowfin tuna to the top lay- 

 ers of the thermocline, with excursions into the mixed 

 layer and to the surface, and a nighttime shallowing 

 into the mixed layer, are consistent with patterns 

 observed by Yonemori (1982) and Carey and Olson 

 (1982). The yellowfin tuna in the current study dis- 

 played a particularly strong adherence to the interface 

 between the surface mixed layer and the thermocline 

 when exhibiting straight-line movements. Similar 

 behavior has been previously noted by Carey and Olson 

 (1982). In the present study, these "traveling" fish 

 would often abruptly alternate between the surface and 

 the thermocline interface, both of which might serve 

 to assist straight-line orientation in what is otherwise 

 a truly three-dimensional habitat. By contrast, fish 

 orienting around FADs showed either little vertical 

 movement or vertical movements that were sinusoidal 

 and largely within the mixed layer. 



The depth and temperature distributions derived 

 from the bigeye tuna tracks are in good agreement with 

 distributional data obtained by analysis of longline 

 fishing success (Saito 1975; Hanamoto 1976, 1987). 

 These fishery data indicate maximum abundance at 

 depths >200 m and in temperatures between 11° and 

 15°C, where the dissolved oxygen is >1 mL/L. How- 

 ever, the tracking data suggest daytime distribution 

 (220-240 m, 14-17°C) is influenced at least as much 

 by temperature as by depth, and none of the catch-rate 

 studies indicate the dramatic nighttime upward shift 

 that the tracking data reveal. 



The average sustained swimming speed derived from 

 the straight sections of the current yellowfin tuna 

 tracks was 4.46 Km/hour (range 3.2-6.5 Km/hour, 

 Table 1), which is similar to the 4.27 Km/hour average 

 (range 2.4-7.8 Km/hour) reported by Carey and Olson 

 (1982) for fish having an average fork length 41% 

 greater than those used in our study. This comparison 

 would suggest that absolute traveling (sustainable) 

 swimming speed for this species does not increase with 

 size, at least for fish between 45 and 100 cm FL. This 

 constancy of swimming speed, regardless of increas- 

 ing fish length, is consistent with the model generated 

 by Magnuson (1973, 1978), which predicts that in- 

 creased weight associated with increasing length is 



