Leis et al. Behavioral ontogeny in larvae and early luvenile Caranx ignobilis 



403 



B 



D 



Figure 1 



Preserved larvae of giant trevally ^Caranx ignobilis) representing the range of sizes and developmental stages included 

 in the present study. Preopercular spination was present in all specimens. The smallest specimen had a few scutes on 

 the peduncle, the 14.5-mm specimen had scales forming elsewhere, and the largest specimen was largely scaled. Sizes 

 are SL (Australian Museum catalogue numbers are provided). lA) 9.6 mm (1.43435-004); (B) 11.7 mm (1.43435-003); 

 (C) 14.5 mm (I. 43435-004); (D) 18.5 mm (1.43435-002). 



flowing over the outlet weir to fill a container of known 

 volume, divided by the cross-sectional area of the cham- 

 ber. The average of five calibrations was used as the 

 flow speed for a given valve angle. The chambers were 

 calibrated each time they were set up. Flow speeds in 

 excess of 50 cm/s could be achieved with this system. 

 The chambers were plumbed into the seawater system 

 of the NMMBA aquaria and laboratory, which provided 

 a continual flow of "fresh" seawater. 



Before measurements of swimming speed and endur- 

 ance were recorded, the larvae were acclimated to any 

 differences in water quality between the holding tank 

 and swimming chamber by gradual addition of seawater 

 from the swimming chamber system. Larvae were placed 

 in a chamber lane and allowed to acclimate for 5 minutes 

 at 1 cm/s. Any larva showing signs of stress during the 

 acclimation period was removed from the experiment and 

 replaced with another individual. Water temperature in 

 the swimming chamber ranged from 27° to 29°C. Two 

 swimming parameters were measured: critical speed 

 (t/|,^i(), which measures maximum swimming speed over 



periods of minutes, and endurance, which measures how 

 long larvae can swim without food or rest. 



For C/pj.j, tests, starting at 1.5 cm/s flow speed was in- 

 creased by approximately 2 cm/s every 5 minutes until 

 the larvae were unable to swim against the flow. The 

 elapsed time when each larva drifted to the downstream 

 mesh was recorded. Critical speed (U^^^^) of larvae was 

 calculated with the equation of Brett (1964): 



U., 



U + U/t, X U,), 



where U = penultimate speed; 



C/, = speed increment (ca. 2 cm/s in the present 



study); 

 t = time swum in the final speed increment; 



and 

 t^ = the time interval for each velocity increment 



(5 min). 



The total time for a critical speed measurement was 

 proportional to the f/^.^^ achieved, and varied from 15 



