402 



Fishery Bulletin 104(3) 



ently moved out of the bay when larger than 40 cm. In 

 South African estuaries, individuals larger than 55 cm 

 were absent (Blaber et al., 1983). The high abundance 

 of C. ignobilis in the NW islands of Hawaii (Sudekum 

 et al., 1991), where estuaries are absent, shows that 

 presence of juveniles in estuaries is facultative. 



Recently, Taiwanese aquaculturists have developed 

 culture methods for C. ignobilis (Yu. 2002), and this 

 enabled us to obtain larvae and small juveniles that we 

 used for laboratory and field observations of behavior. 

 Behavior of early-life history stages of marine fishes can 

 strongly influence both the survival and distribution of 

 these small fishes (Leis and McCormick, 2002). There- 

 fore, an understanding of the behavioral capabilities of 

 these early-life history stages is essential to intelligent 

 management, and to comprehend the demographics and 

 ecology of any marine teleost fish. 



Aside from field studies of the behavior of wild larvae 

 in vertical distribution (e.g., Ahlstrom, 1959; Olivar and 

 Sabates, 1997; Flores-Coto, et al., 2001), behavior has 

 been studied in the early-life history stages of only a 

 few carangid fishes. There has been extensive research 

 on cannibalism, aggressive interactions, and school- 

 ing behaviour by using reared larvae and juveniles 

 of Japanese amberjack (Seriola quinqueradiata) and 

 white trevally (Pseudocaranx dentex) in the labora- 

 tory (e.g., Masuda and Tsukamoto 1996, 1998, 1999; 

 Sakakura and Tsukamoto 1996, 1999). We report in 

 the present study the first observations of the develop- 

 ment of swimming abilities, orientation abilities, and 

 vertical distribution for C. ignobilis, using both in situ 

 methods (Leis et al., 1996) and laboratory swimming 

 chambers (Fisher et al., 2000). These behaviors are 

 important to dispersal, feeding, predator interactions, 

 and survival generally during the early life history of 

 C. ignobilis. 



Materials and methods 



Larvae and juveniles 



Young C. ignobilis were obtained from a commercial 

 aquaculture farm near Kaohsiung, Taiwan. They were 

 identified as C. ignobilis by the farmer with reference 

 to photographs, and the identification was subsequently 

 confirmed by examination of preserved specimens. 

 Caranx ignobilis eggs from an induced spawn were 

 placed in a large outdoor earth pond (approximately 

 20x20xlm), and thus hatched under "natural" condi- 

 tions. Larvae were provided with a "natural" food source 

 (phytoplankton and zooplankton that were resident in 

 the pond). Surface water temperature in the outdoor 

 pond was 32°C when the larvae were collected in May 

 2004. Fish were obtained on two occasions at the same 

 pond from a single cohort and from an unknown number 

 of females; the first collection was at 20 days after hatch- 

 ing (dab) and the second at 24 dah. 



The young fish were placed in oxygenated plastic bags 

 in an insulated box and transported to the laboratory at 



the National Museum of Marine Biology and Aquarium 

 (NMMBA), Renting, Taiwan, about 1 hour by road. In 

 the laboratory the larvae were acclimated in a 40-liter 

 aquarium filled with water from the seawater system 

 at NMMBA. Each aquarium was fitted with an aerator, 

 and kept ca 25°C. Twice daily, the larvae were fed with 

 live, newly hatched brine shrimp (Artemia) nauplii and 

 50% of the total volume of water exchanged with fresh 

 seawater. The aquaria were cleaned daily by suctioning 

 debris off the bottom. 



Reported sizes of larvae are given in standard length 

 (SL, in mm). Ages are reported as days after hatching 

 (dah) and are based on the age reported by the aqua- 

 culturist when the larvae were obtained. The nomen- 

 clature for early life history stages of fishes is complex; 

 there are many different systems of terminology and no 

 consensus on which is the most appropriate. Depending 

 on the nomenclature used, the C. ignobilis individuals 

 we studied (Fig. 1) would be considered larvae, or ju- 

 veniles, or as a mixture of both. In our C. ignobilis, all 

 fin-rays were present in the smallest specimens (8 mm), 

 and scales were present from about 14 mm, yet the 

 preopercular spines that characterizes most carangid 

 larvae (Leis and Carson-Ewart, 2004) were still pres- 

 ent at 19 mm. Unlike the pelagic stages of demersal 

 fishes, C. ignobilis does not undergo a clear ecological 

 transition from pelagic to demersal habitat upon which 

 one might base life history stages. We do not attempt to 

 distinguish between larvae and juveniles, and to avoid 

 awkward phrasing and for simplicity, we refer to the 

 young fish that we studied as larvae on the basis that 

 the largest individuals retained some head spines. We 

 acknowledge that in some taxonomic systems they may 

 be referred by other terms. 



Laboratory observations 



Multilane swimming chambers were used to measure 

 swimming abilities (Stobutzki and Bellwood, 1994). 

 One chamber was used to measure critical speed and 

 a second identical chamber was used for measurements 

 of endurance (Fisher et al., 2000). Both chambers were 

 made of clear plexiglass and had 6 lane-ways, each 30 

 mm wide, 50 mm high, and 180 mm long. A black line 

 across the lid of the chamber provided the larvae with 

 a point of reference for orientation. Aside from the fine 

 mesh (0.5-mm) ends, the chamber design was identical 

 to that of Stobutzki and Bellwood (1994, 1997). 



Even distribution of flow was achieved by a T-piece 

 diffuser in the header portion of the chamber. Turbu- 

 lence in the chamber was minimized by a 40-mm-long 

 section of flow straighteners at the start of each lane. 

 These straighteners also minimized possible boundary 

 layers. Previous measurements have shown that water 

 velocity in the 5 mm area closest to the wall was not 

 significantly different from that in the center of the 

 chamber (Stobutzki and Bellwood, 1997; Stobutzki, 

 1998; Fisher et al., 2000). Water flow speed was con- 

 trolled by turning a calibrated ball valve. Flow rates 

 were calibrated by recording the time taken for water 



