68 



Fishery Bulletin 91(1). 1993 



as 



35 



10 



15 



americaiuis. 



phosis to correlate behavior with 

 physiological (as registered in 

 otolith development) and mor- 

 phological changes. Observations 

 of the larvae were made in hold- 

 ing tanks throughout the period 

 they were reared. Individual lar- 

 vae were also observed in small 

 containers under low magnifica- 

 tion to verify anatomical changes 

 as well as details of small move- 

 ments. 



On the 48th day after hatch- 

 ing, individuals were moved into 

 an observation tank to facilitate 

 observation. An undergravel 

 filter bed was placed in the re- 

 frigerated observation tank to 

 minimize disturbance resulting 

 from maintenance procedures. 

 Temperature, diets, and light- 

 cycle conditions were the same 

 as those for separate tray-reared 

 larvae. Light intensity was 

 higher in the observation tank 

 than in trays because overhead 

 lights were supplemented with tank lights. Fish were 

 observed for lOmin periods twice daily, at 10 a.m. and 

 4 p.m. One fish chosen at random was followed as long 

 as it could be seen; if it moved out of sight, another 

 individual was selected for the remainder of the obser- 

 vation period. Behaviors recorded included swimming 

 (duration, vertical and horizontal direction, body ori- 

 entation in relation to the bottom, and fin usage); feed- 

 ing, both before and after adding food (frequency, loca- 

 tion, sequence of body motions, success); resting 

 (duration, location, body position); and interactions be- 

 tween individuals. Observations were terminated sev- 

 eral weeks after fish had metamorphosed and behav- 

 ior patterns had stabilized (i.e., assumed a typical adult 

 sedentary behavior pattern). 



Results and discussion 



Larval growth rates 



From analysis of 113 preserved larval winter flounder 

 ranging from 2.5 to 9.0mmSL, growth was best de- 

 scribed by a Gompertz-type curve (Fig. 3). Previous 

 uses of the Gompertz growth curve and methodology 

 for fitting the curve are described in Pennington ( 1979) 

 and Bolz & Lough (1988). The variance was stabilized 

 by using the natural log form of the growth equation, 

 and parameters were derived by nonlinear estimation 

 techniques resulting in the relationship: 



20 25 30 35 40 



Estimated Age (d) 



45 



50 



55 



60 



Figure 3 



Gompertz growth curve and equation fitted to plot of standard length and estimated age 

 in days Ino. of otolith increments + 10) for 113 larval winter flounder Pleuronectes 



ln(L) = -0.3469+2. 5329(l-e- 



r-=0.8106, (1) 



where L = standard length in mm, and R = estimated 

 age (increments +10) in days. 



The predicted length of 2.66 mm at yolksac absorp- 

 tion compares favorably with that found by Radtke & 

 Scherer (1982) for wild larvae (2.5 mm). The asymptotic 

 length of 8.9 mm probably delineates mean length at 

 metamorphosis and falls within the range (7-13 mm) 

 given by Fahay (1983). The average growth rate 

 (from Eq. 1) for the period under study was 0.31 mm/d, 

 which is slightly less than that observed in the 1982 

 study using preserved lengths by Radtke & Scherer 

 (0.38 mm/d). 



Shrinkage 



Little shrinkage was observed for larvae 4-35 d old 

 as has been reported by other researchers (Radtke & 

 Waiwood 1980, Theilacker 1980). Fresh lengths were 

 2.8-5. 0mm; preserved lengths, 2. 5-5. 0mm (n=28) with 

 average shrinkage of 4.2<7c (SE=0.6). For 91-112d old 

 flounder, fresh lengths were 5.9-13.8 mm; preserved, 

 5.8-13.6 mm (n=19) with average shrinkage of 8.6% 

 (SE=0.9). Radtke & Scherer (1982) found no shrinkage 

 in small larvae (<4.7mm) and only minimal shrinkage 

 (4%) in older fish. Though we observed slightly larger 

 shrinkage than Radtke & Scherer (1982), the growth 

 rates during the 55 d agree roughly with larval flounder 

 growth rates found in their study. 



