846 



Fishery Bulletin 101(4) 



f 1?- 



E 10- - 



Abudcfduf twfchelii 



enriched i-ADs 

 Y =- 2.2X + s.y 



non-enriched FADs 

 Y = 0.1X-0,1 



0-$ 



00:00 



-*- 



-^ 



■V 



01:00 02:00 03:00 04:00 

 Elapsed time (tih.mm) 



Figure 9 



Changes in the number of fish associated 

 with enriched (solid circles) and nonen- 

 riched (open circles) drifting FADs over time. 

 Enriched FADs showed significantly higher 

 rates of recruitment than did nonenriched 

 FADs; the slopes of the regression lines are 

 significantly different ((=3.05, 2-tailed test, 

 v=6,P=0.02). 



There appears to be insufficient time or stability for such 

 factors as competition or predation to influence the size or 

 diversity of these FAD-associated fish assemblages. De- 

 spite these results, individual fish do remain with a specific 

 FAD for days: On at least five separate occasions associated 

 with some of the other experiments described in this paper, 

 individuals recognizable by scars and bite marks were 

 sighted repeatedly as many as six days after the initial 

 observation (Nelson, unpubl. data). Although the FADs and 

 the associated fishes described in the present study are not 

 directly comparable to FADs and fishes targeted in fisheries- 

 scale operations, these experiments are among the first con- 

 trolled efforts at understanding the effects of disturbance or 

 fishing for FAD-associated assemblages. 



The average assemblage size for all experiments and 

 treatments varied considerably, often significantly, over 

 time (Tables 2-5). Significant interaction effects between 

 sample date and FAD treatments may be indicative of day- 

 to-day recruitment fluctuations, dependent upon recruit- 

 ment variability. A significant interaction may result when 

 these effects are large and are in evidence regardless of the 

 experimental treatment (i.e. occur in concert across treat- 

 ments). Significant sample date effects (and series effects 

 in the fish-removal experiment, Table 2) are likely a result 

 of temporal fluctuations in the numbers of fishes available 

 to recruit to the FADs. 



Note that the two series in the fish-removal experiment 

 differed not only in which FADs were given a particular 

 treatment (positional effects), but also in time — the two 

 scries were necessarily run consecutively, not concurrently. 

 I believe, however, it to be unlikely that positional effects 

 influenced any of the results reported in the present study: 

 treatments were assigned to FADs within the arrays in 



such a way as to ensure that inshore, offshore, or longshore 

 positions were equally weighted among treatments. Signifi- 

 cant series main effects, independent of additional factors, 

 were found only for species richness and diversity (HB) — a 

 result I attribute to changes in the availability of potential 

 recruit species. Temporal patterns of juvenile reef fish re- 

 cruitment are often variable and may be affected by such 

 factors as spawning periodicity (Love et al., 1990), variable 

 predation (Nelson, 2001), or changing physical oceano- 

 graphic processes (Doherty, 1991; Levin, 1994; Kingsford 

 and Finn, 1997). Rountree ( 1989), also, found that the mean 

 numbers of the most abundant species observed around a 

 FAD array off South Carolina varied widely during FAD 

 deployment, albeit over a much longer time period (nearly 

 200 days). Thus, differences in assemblage size and diver- 

 sity over time are not unexpected. 



FAD size had significant, positive effects on assemblage 

 size and species richness. Although tripling the FAD size 

 resulted in a nearly threefold increase in the number of as- 

 sociated fishes (combined species), the response may not be 

 linear. (Note, however, that Rountree (1989) demonstrated 

 that the number of Decapterus punctatus associated with 

 midwater FADs exhibited a significant, positive linear re- 

 sponse to FAD size.) Further research will be necessary to 

 resolve the effect of FAD size on numbers of aggregating 

 fishes. Also of interest is the significant increase in species 

 richness attributable to increased FAD size. Bortone et al. 

 (1977) suggested that species diversity may be a function 

 of "clump size" for Sargassum-associated fish assemblag- 

 es, and Moser et al. (1998) found greater numbers offish 

 species under large ( 10-20 m diameter) mats of floating 

 Sargassum than they did under smaller clumps (<1 m 

 diameter) or in open water. However, the changes in spe- 

 cies richness from this experiment could well be an effect 

 of assemblage size; treatment effects on species diversity 

 measured using the Brillouin index (HB) were marginally 

 nonsignificant (Table 3, P=0.07). Significant sample date 

 differences in treatment and evenness are due to large 

 fluctuations in the abundance of the dominant species, 

 Abudefduf troschelii, ranging at the triple-size FADs from 

 1 to 55 individuals over the course of 11 days. 



Fishes were five times more numerous on average at 

 fouled FADs than they were at comparable FADs lacking 

 fouling organisms, but measures of diversity showed no 

 significant treatment effect (Table 4). There was a signifi- 

 cant interaction between treatment and sample date for 

 the present experiment (Table 4) that may have been due 

 to fluctuations in assemblage sizes among sample dates 

 across both FAD treatments. The species composition of 

 these assemblages was similar to that of other experi- 

 ments, except that Elagatis bipinnulata were regularly 

 observed: Abudefduf troschelii were the dominant species 

 by abundance, followed by E. bipinnulata, and Mugil sp. 

 All were small, young-of-the-year fishes (the largest E. bi- 

 pinnulata individuals reached approximately 80 mm SL) 

 and seemed not to be feeding on the larger invertebrates 

 forming much of the colonizing community. During casual 

 observations of FAD-associated fishes, I observed fish feed- 

 ing on plankton carried past the FADs, but no physical 

 contact with the FAD or fouling organisms. Ibrahim et al. 



