Nelson: Fad characteristics and associated fish assemblages 



837 



Fig. 3) and in shallow water (14-25 m). The FAD 

 treatments were not assigned randomly to FAD 

 positions; instead, I assigned treatments uni- 

 formly across the FAD array because the total 

 number of FADs was relatively low (8-10) and, 

 given the small number of experimental units, 

 in order to reduce the possibility that the results 

 be confounded by positional effects. 



The anchored FADs were spaced approximately 

 100 m apart. The maximum horizontal underwa- 

 ter visibility measured was 27 m, and typically 

 averaged much less. Assuming that vision was 

 the principal means by which fishes located 

 these objects, it is therefore highly unlikely that 

 fishes treated the FAD array as a single "object" 

 or moved from FAD to FAD within the array. I 

 believe that it was unlikely that any fish trans- 

 ferred from one FAD to another for the following 

 reasons. 1) The horizontal underwater visibility 

 was always much less than the distance between 

 FADs. 2) Observations suggest that short-term, 

 daytime fidelity was high; once a fish associated 

 with a floating object, it was unlikely to leave that 

 object during the day (Nelson, 1999). 3) Crossing 

 an open stretch of water for another floating ob- 

 ject (presumably within detection range) entails 

 a potential risk for a fish. Moser et al. (1998) did 

 note that the larger juveniles and adult fishes 

 associated with floating Sargassum showed little 

 apparent fidelity to this habitat and would move 

 between their research boats, floating observation 

 equipment, and the Sargassum habitat. However, 

 the fishes that were observed to move between 

 these floating objects were juvenile carangids be- 

 tween 10 and 20 cm in length (Moser et al., 1998), 

 whereas the fishes in the present study were gen- 

 erally much smaller and presumably less vagile. 



Longshore currents ran roughly west to east 

 through the experimental area, and rarely in the 

 reverse (Fig. 3). I recorded an estimate of cur- 

 rent direction using an underwater compass and 

 the angle of the FAD anchor lines. This estimate 

 represented the sum of the forces due to windage 

 on the FAD buoys and currents. 



Censusing FADS 



FAD-associated fish assemblages were censused 

 by direct visual observation by divers using mask 

 and snorkel. Most other studies offish assemblages 

 associated with floating material have employed 

 nets or quantitative fishing methods for sampling 

 purposes (e.g. Kojima, 1960; Dooley, 1972; ICings- 

 ford, 1992, 1995), but Hunter and Mitchell (1968) 

 compared data from net captures, automated 

 photography and direct visual observations and 

 found that the visual observations agreed well 

 with the other methods, and provided behavioral 

 information not available with the other methods. 

 In my study, a FAD was approached by swimming 



Fish Aggregation Device 

 (FAD) 



2 kg line weight 



cast concrete with 6.25 mm 



polypropylene loop 



FAD line & anchor line 

 9 mm blue polypropylen 



to bottom & 25 kg 

 concrete anchor 



Figure 2 



Design of detachable fish aggregating devices (FADs) used in the 

 present study. 



— '-, ^Laboratorio A^fiotines 



long-shore Currents 



North 



\to Los Frailles 

 i9-12km 



rocky coast 

 sand beach 



experimental FAD array 

 (general location) 



t 



drifting FAD experinnents 

 conducted in this area 



Figure 3 



Coastline and general location of experiments and observations. The 

 laboratory iLaboratorio Achotines) is located at approximately T'S'N lat, 

 SOnCW long. 



