404 



Fishery Bulletin 104(3) 



Table 1 



Location-specific relationships between in situ speed (cm/s) and size (mm SL) in Caranx ignobilis larvae. NA=not applicable. 



Location 



Regression 



Wan Li Tong 

 Her Chen 



Nan Wan Bay South 

 Nan Wan Bay North 



Spd = 1.69SL - 6.2 

 Spd = 2.55SL - 12.2 

 Spd = 0.12SL + 4.9 

 Spd = 2.05SL - 17.9 



minutes for the slowest individual to 82 minutes for 

 the fastest. Given the endurance of which these larvae 

 are capable (see "Results" section), it is unlikely that 

 the larvae would have become fatigued over such time 

 intervals. 



For endurance tests, a constant speed of 10 cm/s was 

 used. Larvae were swum until they fatigued, which 

 was defined as the time when a larva could no longer 

 hold its position against the current and drifted onto 

 the downstream mesh. During daylight hours, larvae 

 were observed regularly and the exact time of fatigue 

 (=swimming duration) was recorded. If a larva fatigued 

 when it was not being observed, the time at fatigue was 

 estimated as the midpoint between the time when the 

 larva was last seen swimming and the time when it was 

 found no longer swimming. Chambers were set up under 

 cover so that they were shaded throughout the day, and 

 a fluorescent light was used for illumination at night. 

 The actual endurance measurements of time swum were 

 converted to distance swum, by using the flow speed 

 data, and were reported as kilometres swum. 



Fatigued larvae from both experiments were removed 

 from the chamber, euthanized and fixed in Bouin's so- 

 lution for one hour, then placed in 70% alcohol and 

 stored. All preserved larvae were later examined under 

 a dissection microscope to determine standard length 

 (SL) and state of notochord flexion. Total lateral area 

 (TA) and propulsive area (PA) of larvae (Fisher et al., 

 2000) were measured by using Scion Image for Windows 

 (Beta 4.02, Scion Corporation, Frederick, MD). PA is TA 

 minus the head and gut. 



In situ observations 



Four sites were used for in situ observations of C. ignobi- 

 lis in the South China Sea. at the southern tip of Taiwan 

 (ca. 22"N, 121°E). Two study sites were used in Nan Wan 

 Bay (21 May 2004) where the depth range was 14-23 m 

 and another two were used on the west coast just off the 

 peninsula that delineates the west side of Nan Wan Bay 

 in the vicinity of Wan Li Tong and Her Chen (14 May 

 2004) at a depth range of 17-31 m (Table 1). At each site, 

 observations were made at least 50 m offshore, and all 

 observations were conducted in the morning. 



Larvae were transported from the laboratory to the 

 release sites in covered buckets fitted with a battery-op- 

 erated aerator. Ambient seawater was gradually added 



to the buckets to allow the larvae to acclimatize to the 

 surrounding water conditions. Then, 50% of the water 

 in the bucket was exchanged for "fresh" seawater every 

 hour. The behavior of the larvae was observed following 

 the procedures of Leis et al. (1996) and Leis and Car- 

 son-Ewart (1997, 1998). Two scuba divers descended to 

 a depth of 5 m where the observer diver released a larva 

 from a small container. Once the larva chose its initial 

 trajectory, the two divers followed. The observer diver's 

 sole job was to follow the larva while the second diver, 

 following the observer, recorded data. The direction the 

 observer diver was facing when he or she released the 

 larva was chosen at random. Each larva was used only 

 once and, where possible, was recaptured at the end of 

 the observation period and preserved. The size of each 

 larva was estimated, with the aid of a ruler, before re- 

 lease. For recaptured larvae, the estimated and actual 

 sizes were regressed, and this relationship was used 

 to calculate the actual size of the five individuals that 

 were not recaptured of the 24 released. Water column 

 depth was measured by the depth sounder on the sup- 

 port boat at the start of each larval release. 



Swimming speed, depth, and swimming direction 

 were measured In situ. We attempted to observe each 

 larva for 10 minutes, taking measurements of swim- 

 ming direction and depth, using a dive compass and 

 computer, respectively, every 30 seconds. Speed was 

 calculated from distance traveled as measured by a 

 calibrated flowmeter over the full period of observa- 

 tion (Leis and Carson-Ewart, 1997). Larvae were not 

 followed deeper than 15-18 m on some dives for safety 

 reasons; therefore observations on some individuals 

 were curtailed. Twenty-four larvae were observed, 

 and observations on 18 of the larvae lasted the full 10 

 minutes. Five larvae swam monotonically downward to 

 below our safety depth from 3 to 6 minutes, whereas 

 one initially ascended, before descending after 5 min- 

 utes and exceeding the safety depth after 8 minutes. 



Data analysis 



To determine the best predictor of performance, values of 

 critical swimming speed and endurance were regressed 

 against SL, TA, and PA by using linear, logarithmic, 

 power, and exponential models. The model with the 

 greatest r^ (coefficient of determination) was used as 

 the best description. 



