David et al.: Age and growth in juvenile Sciaenops ocellatus 



513 



Table 1 



Comparison of ages derived from sagittal sections of 

 each otolith type in known age (46 days), hatchery- 

 reared red drum, Sciaenops ocellatus. Values are 

 reported for mean age, standard deviation, and co- 

 efficient of variation. 



Otolith 



N 



Mean age (d) 



SD 



CV 



Asteriseus 



Sagitta 



Lapillus 



50 

 50 

 50 



39.7 

 25.0 

 21.0 



3.2 



4.6 

 5.5 



8.0 

 18.6 

 26.3 



determined by examination of the age series of lar- 

 vae, indicating that rings were indeed formed daily, 

 and all rings were visible. Therefore, we assumed 

 that rings were also formed daily in the lapillus and 

 that underestimation of age was due to the inability 

 to observe rings of low contrast in the opaque nuclear 

 area rather than to other possibilities that have been 

 reported, i.e. nondaily ring formation due to poor 

 growth in herring, Clupea harengus, and turbot, 

 Scophthalmus maximus (Geffen, 1982), and to ring 

 spacing below the resolution limit of light micros- 

 copy in striped bass, Morone saxatilis, under subop- 

 timal feeding regimes (Jones and Brothers, 1987). 



Peters and McMichael ( 1987 ) had difficulty distin- 

 guishing the innermost rings in sagittae of some ju- 

 veniles, and they developed an ageing method that 

 did not require counting these rings. This method 

 utilized transverse sections of larval sagittae in which 

 rings were clearly visible to determine the distance 

 from the primordium to the tenth ring. Subsequent 

 juvenile ring counts were initiated at this distance 

 away from the primordium. We also encountered dif- 

 ficulty in detecting all the rings near primordia in 

 sagittae (and lapilli as well) due to the opacity of the 

 nuclear region of the otoliths, but all rings were usu- 

 ally clearly visible in asterisci. Peters and McMichael 

 (1987) made relatively accurate ring counts beyond 

 the tenth ring on sagittae of 21-day-old known-age 

 fish. However, we were not able to duplicate their 

 success using our technique, because our counts 

 underestimated the true age of 46-day-old fish by 21 

 days with a coefficient of variation of 18.6%. Duplicat- 

 ing their glycerin soaking technique did not improve 

 our ring detection capability. The principle difference 

 between our method and that of Peters and McMichael 

 ( 1987) was the sectioning plane; they used the trans- 

 verse plane, whereas we used the sagittal plane. 



Because the otoliths were birefringent, they were 

 easily observed within larval fish when illuminated 

 with transmitted light between crossed polaroids. 

 Sagittae and lapilli were easily seen in fish as small 

 as 1.3 mm SL, whereas asterisci were observed in 



all fish greater than 3.0 mm SL. Comyns et al. ( 1989) 

 detected two rings in sagittae and lapilli of larvae 2 

 days after fertilization ( 1 day posthatch), demonstrat- 

 ing that sagittae and lapilli are present at hatching; 

 according to their growth curve, fish 3.0 mm SL were 

 approximately 6 days old. Whereas the size at hatch 

 was relatively constant, growth varied considerably 

 with temperature, but the variation was not signifi- 

 cant until the fish attain 4.0 mm SL (Comyns et al., 

 1989). Changes in growth rates due to water tem- 

 perature are therefore not likely to have a signifi- 

 cant effect on our estimate of age at asteriseus for- 

 mation. Others have reported asteriseus formation 

 at similar ages, e.g. at age 6 days in the Japanese 

 eel, Anguilla japonica (Umezawa et al., 1989). 



The asteriseus provided the most accurate estimate 

 of age for juvenile red drum because it underesti- 

 mated true age by only 6 days compared with 21 days 

 for the sagitta and 25 days for the lapillus. The 

 asteriseus ages were also the most precise, because 

 the coefficient of variation was only 8.0% compared 

 with 18.6 and 26.3% for the sagitta and lapillus re- 

 spectively. Most of the variance in age estimates for 

 all three otoliths was caused by the inability to re- 

 solve rings near the primordium and at the margin. 

 Rings in the mid portion of asteriseus sections were 

 consistent in shape, increment width, and clarity; a 

 dominant, consistently identifiable, first ring was 

 visible at the edge of the primordium. 



The sagitta provided the next most accurate and 

 precise estimates. Ages were underestimated by 21 

 days on average, and rings near the primordia were 

 difficult to distinguish, resulting in a coefficient of 

 variation more than a factor of two higher than that 

 of the asteriseus. Because sagittae were present at 

 hatching and early rings were detectable in smaller 

 and younger fish (Peters and McMichael, 1987; 

 Comyns et al., 1989), better accuracy and precision 

 of ageing using sagittae may be possible with im- 

 provements in preparation and processing of the 

 otolith. However, the Spurr mounting technique 

 (Haake et al., 1982) used by Peters and McMichael 

 ( 1987) and Comyns et al. ( 1989) can require several 

 days for proper dehydration, curing, sectioning, and 

 polishing whereas our polymer method required less 

 than one hour to produce ground and polished slides 

 from whole fish, thus allowing considerably more fish 

 to be processed in comparable time periods. 



Ages estimated from the lapillus were least accu- 

 rate and precise because of poor clarity of rings near 

 the primordia and because of the formation and fu- 

 sion of numerous accessory primordia. This fusion 

 of accessory primordia resulted in superimposition 

 of rings near the margin and the presence of several 

 planes of growth being visible in the same focal plane. 



