412 



Fishery Bulletin 104(3) 



mon in larvae of other species (Leis and Carson-Ewart, 

 1999, 2002). We observed one individual penetrate a 

 thermocline, indicating not only a tolerance for rapid 

 temperature change, but also, that the thermocline is 

 not a barrier to vertical movement for C. ignobllis, as 

 has been proposed sometimes for larvae of other species 

 (e.g.. Gray and Kingsford, 2003). 



We used reared larvae in our study and it is possible 

 that the behavior we observed may have differed from 

 that of wild individuals. Results of studies comparing 

 wild and reared individuals have been inconsistent. 

 Some studies have shown large differences in behavior 

 of reared and wild fishes, the most marked of which 

 involve learned interactions with predators (Brown 

 and Laland, 2001). Differences in swimming perfor- 

 mance (in both directions) have been documented in 

 some studies, but not others, or have been present at 

 some developmental stages but not others (e.g., Blax- 

 ter, 1975; Dunmall and Shreer, 200.3; Smith and Fui- 

 man, 2004). Even if differences in predator-related 

 behavior exist, swimming and orientation are pre- 

 sumably less dependent on experience and learning, 

 and there is less reason to expect them to differ as 

 well. The possibility that reared larvae of C. ignobilis 

 may behave differently from wild larvae cannot be 

 entirely dismissed, but the fact that C. ignobilis larvae 

 were reared in conditions approaching those found in 

 the field (but with the absence of predators) — kept in 

 large outside ponds at an aquaculture farm and fed 

 a natural assemblage of zooplankton — would presum- 

 ably make large differences in behavior with their 

 wild counterparts less likely. Unfortunately, we did 

 not have access to wild C. ignobilis, and therefore we 

 could not make direct tests; moreover, little is known 

 of behavior in other carangid larvae upon which we 

 might base comparisons. 



Caranx ignobilis apparently schools when young (My- 

 ers, 1999), and the size range of larvae we studied 

 overlapped with that at which schooling begins in other 

 carangid species (Masuda and Tsukamoto, 1998). We 

 did not attempt to observe schooling in C. ignobilis, but 

 it is conceivable that some of the aspects of behavior 

 may differ between individuals that are schooling and 

 those that are solitary. 



We observed strong ontogenetic changes in behavior 

 of C. ignobilis across a size range of 8 to 18 mm SL. 

 At any size, there was usually a wide range of perfor- 

 mances present, and the best performers were able to 

 swim much better than average performers. Given the 

 very high mortality rates experienced by marine fish 

 larvae, the average larva at any given size is unlikely to 

 survive (Gushing, 1990), and it may be only the excep- 

 tional performers that do survive. Therefore, it may be 

 more appropriate to use values for the best performers 

 than for average performers in models of dispersal or 

 survival. Increases in swimming abilities with growth 

 were most marked, and significant swimming abilities 

 would probably be present, in larvae as small as 5 mm 

 SL based on extrapolation of performance at size rela- 

 tionships. Ontogenetic changes in vertical distribution 



behavior were more complex and difficult to interpret, 

 but depth selection seemed to be more variable (both 

 within and among individuals) in larger larvae. We 

 found that orientation behavior of small C. ignobilis was 

 already developed and did not increase ontogenetically, 

 although the direction they swam differed between 

 locations. Behavior in small C. ignobilis was complex 

 and swimming abilities were well developed. These 

 behaviors are capable of strongly influencing survival 

 and dispersal and show that the larvae of some pelagic 

 fishes have behavioral capabilities similar to those re- 

 cently documented for larvae of demersal reef fishes 

 (Leis and McCormick, 2002). 



Acknowledgments 



We thank the Director of NMMBA, Lee-Shing Fang, 

 for the opportunity to work at NMMBA and his staff 

 for their excellent cooperation, particularly Chao-Yuan 

 Chung for assistance in the wet laboratory. Our work 

 could not have proceeded without the assistance of Colin 

 Wen and Kun-Ping Kan in all phases of the work. Li- 

 Hua Chao generously introduced us to many Taiwan- 

 ese aquaculturists and spent many hours helping us 

 obtain larvae. Mark Brown assisted ably, especially with 

 laboratory work. Ray-Ming Chen, our dive boat skipper, 

 made possible our field work, and Bill Watson helped 

 with literature. This research was supported by an ARC 

 Discovery Grant (DP0345876), and a DST International 

 Science Linkages Grant (IAP-IST-CG03-0043) to JML, 

 and by the Australian Museum. Reiji Masuda made 

 constructive comments on the manuscript. 



Literature cited 



Ahlstrom, E. H. 



1959. Vertical distribution of pelagic fish eggs and 

 larvae off California and Baja California. Fish. Bull. 

 60:107-146. 

 Batschelet, E. 



1981. Circular statistics in biology, 371 p. Academic 

 Press, London. 

 Blaber, S. J. M., and D. P. Cyprus. 



1983. The biology of the Carangidae (Teleostei) in Natal 

 estuaries. J. Fish. Biol. 22:173-188. 

 Blaxter, J. H. S. 



1975. Reared and wild fish — how do they compare? Proc. 

 10th Europ. Symp. Mar. Biol. 1:11-26. 

 Brett, J. R. 



1964. The respiratory metabolism and swimming perfor- 

 mance of young sockeye salmon. J. Fish. Res. Board 

 Can. 21:1183-1226. 

 Brown, C and K. Laland. 



2001. Social learning and life skills training for hatchery 

 reared fish. J. Fish. Biol. 59:471-493. 

 Clark, D. L., J. M. Leis, A. C. Hay, and T. Trnski. 



2005. Swimming ontogeny of larvae of four temperate 

 marine fishes. Mar. Ecol. Prog. Ser. 292: 287-300. 

 Gushing, D. H. 



1990. Plankton production and year class strength in 



