Leis et a\ ; Behavioral ontogeny in larvae and early juvenile Caranx ignobilis 



411 



periods at speeds that would make them effective swim- 

 mers in many situations. According to the relationship 

 in Figure 4, C. ignobilis larvae of about 9 mm could 

 swim without food or rest for about 13 hours (5 km at 

 10 cm/s) at a speed that is within their observed ca- 

 pabilities in situ: this ability increases markedly with 

 growth. Clearly, C. ignobilis larvae from a size of 1 cm 

 or less are capable of maintaining speeds of similar 

 magnitude to mean ambient currents for periods of time 

 long enough to strongly influence their dispersal. 



Knowledge of vertical distribution of fishes is impor- 

 tant because many things vary vertically in the ocean, 

 including current velocity, food concentrations, and 

 predators, all of which can strongly influence survival 

 and dispersal. For these vertically varying factors, the 

 conditions that larval and small juvenile fishes actually 

 encounter are fully under their control because swim- 

 ming abilities and sensory abilities capable of determin- 

 ing vertical position in the water column develop at a 

 very small size. The C. ignobilis we observed in the 

 ocean had considerable control over their vertical distri- 

 bution, and the depths selected changed with size. Ma- 

 suda and Tsukamoto (1996) noted ontogenetic changes 

 in preferred light intensity in larvae of P. dentex, and 

 it is possible that such changes influence the selection 

 of depth in larval C. ignobilis. Two notes of caution are 

 necessary in interpreting the depth-selection behavior 

 and ontogenetic changes in this behavior. The larg- 

 est larvae were primarily observed at one site, and, 

 it is possible that vertical-distribution behavior varies 

 among areas (Leis, 2004). Secondly, the largest larvae 

 tended to swim to below our safety depth and we there- 

 fore do not know if they continued to descend, if they 

 subsequently ascended (in accord with the oscillatory 

 behavior observed in a number of larger individuals), or 

 if they leveled off In a similar experiment, much larger 

 (60-140 mm) Pseudocaranx dentex juveniles initially 

 descended several meters before ascending to a much 

 shallower "preferred depth," usually within 60 seconds 

 (Kuwada et al., 2000). We did not observe such ascents 

 following an initial descent, but we cannot rule out the 

 possibility that they occur for individuals that swam 

 below our safe diving depth (Fig. 5D). 



Orientation in the pelagic environment is difficult 

 for fish larvae because of the movement of the water 

 column and the scarcity of reference points in a mov- 

 ing pelagic environment. Orientation is, however, nec- 

 essary for the arrival of late-stage larvae at nursery 

 habitat and can greatly influence dispersal trajectories 

 (Shanks, 1995; Leis and Carson-Ewart, 2003). Two- 

 thirds of the C. ignobilis larvae that we observed swam 

 directionally, as opposed to swimming randomly. This 

 is a somewhat smaller percentage than the 80-100% 

 for directional individuals at the settlement stage for 

 demersal coral reef fishes of similar size (Leis and Car- 

 son-Ewart, 2003). Neither the proportion of individual 

 C. ignobilis that swam directionally, nor the precision 

 of their directional swimming, increased with growth. 

 The lack of ontogenetic change in individual orienta- 

 tion — in contrast to speed, endurance, and vertical dis- 



tribution — indicates that orientation ability develops 

 early (at less than 8 mm SL) in C. ignobilis larvae. Of 

 course, orientation may improve in individuals larger 

 than those we studied. 



The offshore swimming direction of C. ignobilis lar- 

 vae found off the west coast is similar to behavior re- 

 ported for larvae of some other species at other locations 

 (e.g., Leis et al., 1996; Leis and Carson-Ewart, 2003), 

 but could, in fact, be orientation to the west, rather 

 than offshore. Observations off the east coast would 

 be required to separate the possibilities, and any such 

 test would need to take into account the possibility of 

 temporal factors. The lack of any directionality in Nan 

 Wan Bay is possibly a result of the "U" shape of the 

 bay. Such a bay, where "offshore" constitutes less than 

 half of the possible swimming directions, and where 

 west is onshore, may present challenges to orientation 

 that are greater than those present off an open coast. 

 Alternatively, a bay may offer characteristics that would 

 induce larvae to remain, rather than swim away as they 

 did off the more open coast. Or, older larvae may prefer 

 inshore environments, whereas younger ones do not. 

 The important points, however, are that larvae of 8-15 

 mm SL demonstrated orientated swimming in the field 

 and that smaller larvae swam in the same direction 

 as larger ones. We can offer only speculation as to the 

 types of cues that C. ignobilis may use to orient their 

 swimming, and refer the reader to other sources for this 

 information (e.g., Kingsford et al., 2002; Leis and Mc- 

 Cormick, 2002). Our observations on C. ignobilis sup- 

 port other research showing that orientation behavior in 

 fish larvae can be location-dependent (Leis et al., 1996; 

 Stobutzki and Bellwood, 1998; Leis and Carson-Ewart, 

 2003), and if so, this feature add a further complication 

 for attempts to model dispersal. 



Our unplanned in situ observations of behavior shed 

 light on little-known aspects of the early life history 

 of C. ignobilis, which are otherwise difficult to study. 

 Feeding activity by 8-18 mm larvae of C. ignobilis was 

 common while they were swimming, as has been found 

 with other species (Leis and Carson-Ewart, 1998). The 

 only larva to encounter a large pelagic fish reacted in 

 the same way as that reported for other species, namely 

 with the apparent goal of moving away from the poten- 

 tial large predator before the small fish could be readily 

 detected (Leis and Carson-Ewart, 2001). Such behavior 

 in a reared individual with no experience with piscivo- 

 rous fishes indicates that the behavior is not learned. 

 Many carangids associate with jellyfish medusae when 

 small (Shojima, 1962); therefore the brief association we 

 observed between a jellyfish and a 10-mm C. ignobilis 

 was not unexpected. Caranx ignobilis is considered to 

 be reef-associated during some periods of its life history, 

 but individuals of 9-13 mm do not seem to be inclined 

 to associate with coral reefs. Unlike the larvae of many 

 demersal reef fishes (Leis and Carson-Ewart, 2002), 

 the larval C. ignobilis that we observed made no at- 

 tempt to settle and showed no interest in coral reefs. If 

 anything, C. ignobilis moved away from the coral reefs 

 they encountered, although this behavior is not uncom- 



