512 



Fishery Bulletin 92(3). 1994 



otolith primordia were well defined in sagittae, 

 which, along with a uniform axis of growth, resulted 

 in a more consistent axis of measurement. The lapil- 

 lus and asteriscus were more circular in shape than 

 the sagitta and had a less well-defined anterior ros- 

 trum. The lapillus also formed accessory primordia, 

 which were reflected in an irregular perimeter. Al- 

 though accessory primordia were not observed in 

 asterisci, the axis of growth shifted with develop- 

 ment, resulting in diameter measurements through 

 the primordia that did not necessarily reflect the 



maximum otolith diameter. These features resulted 

 in greater variability of diameter measurements for 

 asterisci and lapilli than for sagittae. 



The accuracy and precision of age estimates var- 

 ied among otoliths. Fish used to estimate daily ages 

 were collected 46 days after hatching. Although in- 

 crements were present in all three otoliths, the defi- 

 nition and resolution of the rings differed, especially 

 near the primordia. The mean ring counts, standard 

 deviations, and coefficients of variation are presented 

 in Table 1. All increment widths exceeded the limit 

 of resolution for light microscopy by more 

 than an order of magnitude (Jones and 

 Brothers, 1987; David and Paul, 1989). 

 All ring counts substantially underesti- 

 mated the age of the 46-day-old hatch- 

 ery-reared fish. The asteriscus underes- 

 timated the true age by 6 days, while 

 sagittae and lapilli underestimated age 

 by 21 days and 25 days, respectively. To 

 obtain realistic age estimates using 

 asterisci, ring counts were adjusted be- 

 cause that otolith is not formed at hatch- 

 ing, but age estimates from ring counts 

 in lapilli and sagittae were erroneously 

 low because all rings could not be ob- 

 served and counted because of poor con- 

 trast in the nuclear region. In addition 

 to providing the most accurate age, coef- 

 ficients of variation indicated that the 

 asteriscus was also the most precise in- 

 dicator of true age. 



Discussion 



Shrinkage of larvae due to preservation 

 has been found to be significant (Blaxter, 

 1971; Theilacker and Dorsey, 1980; Hay, 

 1982; Brothers et al., 1983; and Leak, 

 1986). We did not adjust length measure- 

 ments for shrinkage due to preservation 

 because most specimens were juveniles 

 (15-50 mm SL), and because we imme- 

 diately fixed specimens in ethanol which 

 has been shown to minimize the problem 

 of shrinkage (Radke, 1989). 



Daily increment formation has been 

 validated in sagittae by using laboratory- 

 reared red drum that were up to 21 days 

 posthatch (Peters and McMichael, 1987). 

 The 6-day underestimate in age deter- 

 mined from the asteriscus of known-age 

 i hatchery-reared) fish corresponded well 

 with age at otolith formation (6-7 days) 



