Nelson: Fad characteristics and associated fish assemblages 



847 



(1996) reported that none of the gut contents from FAD- 

 associated fishes included sessile organisms found on their 

 FADs (fish size ranges included specimens 8-14, 15-99, and 

 >100 mm SL — the first two size categories are comparable 

 to the fishes in the present study). Larger, piscivorous fishes 

 do feed at least occassionally on smaller fishes associated 

 with floating objects (Gooding and Magnuson, 1967), but 

 published gut content studies are conflicting. Some sug- 

 gest that piscivorous species that associate with flotsam 

 rely on other sources of food (e.g. Gooding and Magnuson, 

 1967; Hunter and Mitchell, 1967; Brock, 1985), while others 

 suggest that flotsam- (or algae-) associated fishes form an 

 important food resource for these larger piscivorous fishes 

 (Dooley, 1972; Manooch et al., 1984; Coston-Clements et al, 

 1991). Morgan et al. (1985) noted the occurrence of at least 

 two members of the Sargosswrn -associated invertebrate 

 fauna among the stomach contents of several species of 

 pelagic fishes. From the perspective of flotsam- or FAD-as- 

 sociated fishes, opportunistic predation by piscivores that 

 do not associate with FADs may be more important than 

 predation by other members of the assemblage. Additional 

 gut content data from juvenile and nonpiscivorous fishes 

 are sorely lacking. I address possible explanations for the 

 results of the present study below. 



I recorded no significant treatment effect attributable 

 to differences between FADs with artificial fish, FADs de- 

 ployed with artificial-fish-sized weights, or control FADs. 

 I attribute significant sample date effects to day-to-day 

 changes in constituent individuals and the fluctuating 

 availability of potential recruits. Numbers of £. bipin- 

 nulata were strikingly constant across treatments and 

 sample dates in this experiment (Table 5, Fig. 8) and in 

 the fouling experiment (Fig. 7) and seemed to indicate an 

 apparently unusual characteristic of this species — individ- 

 uals remaining associated with a given FAD for multiple 

 days. Although the experiment was intended to distinguish 

 between FADs with prior recruits versus FADs without 

 prior recruits, the lack of a significant treatment effect does 

 not negate the possibility that potential recruits would 

 distinguish between occupied and unoccupied FADs. The 

 painted artificial fish and lead weights clearly lacked many 

 attributes of living fish. However, comparable numbers of 

 recruits found at all treatments suggest that a change in 

 the structural complexity of the FADs did not affect assem- 

 blage size or diversity. Although the addition of four small 

 lead weights (artificial fish were painted and oriented 

 differently but were still lead weights) did not appear to 

 increase appreciably the visible surface area of those FADs, 

 the subsequent experiment with live fishes instead of ar- 

 tificial fish had a dramatic effect on recruitment; therefore 

 sizeable changes in the physical size of a FAD may be nec- 

 essary to yield a response in fish recruitment. The potential 

 roles of structural complexity and orientation of FADs will 

 be informative areas for future research. Past investiga- 

 tions in these areas (e.g. Hunter and Mitchell, 1968; Klima 

 and Wickham, 1971; Wickham et al., 1973) have provided a 

 useful beginning, but more work is needed. 



Although sample sizes were small, the presence of prior 

 Abudefduf troschelii "recruits" (enriched FADs) had a sig- 

 nificant effect on patterns of subsequent recruitment; this 



effect contrasted sharply with FADs lacking fish at the 

 start of this experiment (nonenriched FADs). For this spe- 

 cies, these results point to a social aspect to these aggrega- 

 tions, and sociality may also be involved in the recruitment 

 of other species, particularly the schooling fishes Caranx 

 spp., Polydactylus approximans, and Mugil spp., as sug- 

 gested for some scombrids (e.g. Dagorn and Freon, 1999). 

 The addition of fishes below a FAD may increase recruit- 

 ment rates by rendering the object more visible, although 

 the artificial fish experiment indicated that simply adding 

 fish-size objects beneath a FAD does not affect recruitment. 

 Comparisons between these two experiments are tenuous, 

 however, because the artificial fish experiment employed 

 anchored FADs observed over a period of days, whereas 

 the enriched FAD experiment used drifting FADs observed 

 over a course of hours. 



Why do FAD size, the presence of a fouling community, 

 and the actual presence of prior recruits at a FAD each 

 have the effect of increasing the size and, possibly, the di- 

 versity of FAD-associated assemblages of juvenile fishes? 

 The simplest explanation is that these factors contribute 

 to the target strength of the object, increasing the visual, 

 olfactory, or auditory stimulus (or some combination) of 

 the floating object. Larger objects should be easier to find, 

 especially if potential recruits rely on vision to explore 

 their environment. Kellison and Sedberry (1998) found 

 that the fishes associated with mid-water floating struc- 

 tures that were tethered to an artificial reef decreased 

 in abundance over time (193 days), and suggested that 

 the loss in buoyancy associated with the development of 

 a fouling community may have reduced the effective size 

 of these floating objects, accounting for fewer associated 

 fishes (see also Hunter and Mitchell, 1968; Rountree, 1989). 

 To account for the positive effects of a fouling community 

 observed in the present study, it seems reasonable to sup- 

 pose that fouling organisms may be detected by olfactory 

 means; Sweatman ( 1988) has shown that some larval fishes 

 use olfactory cues for settlement on reefs. Further experi- 

 ments, for example experiments controlling for FAD size, 

 odor cues, and visibility of the FADs, are needed to deter- 

 mine why some of these factors exhibit these effects. 



Future research on the role of flotsam as shelter from 

 predators and as a conveyance to suitable habitat could 

 yield evolutionary explanations for the attraction to float- 

 ing objects. For these small fishes, such objects likely repre- 

 sent a shelter from predators (Mitchell and Hunter, 1970). 

 Some species do respond to the approach of an observer by 

 positioning themselves so that the FAD is between them 

 and the observer. Particularly during daylight and crepus- 

 cular hours when visually-oriented predators are most 

 active, flotsam may offer refuge in a habitat where there 

 is little alternative refuge. During the day, when onshore 

 winds drive drifting objects towards shallow water, flotsam 

 and drift algae, unlike anchored FADs, may also offer a 

 comparatively safe conveyance to more suitable habitat. 

 Thus, there may be adaptive advantages for juvenile reef 

 fishes in associating with floating objects. 



Although the juvenile fishes associated with the FADs 

 used in the present study are not of interest to any fishery, 

 the patterns observed from them may be relevant to FADs 



