Ralston et al.: Accuracy of age estimates for Sebastes jordani 



9! 



codes were as follows: 1 = otolith was not readable; 

 2 = low confidence in age estimate, changing the fo- 

 cal plane may have resulted in a different increment 

 count; 3 = moderate level of confidence, preextrusion 

 increments sometimes evident, extrusion check 

 sometimes at an unusually large or small otolith ra- 

 dius; 4 = high confidence in age estimate; and 5 = 

 otolith exhibits remarkable clarity, no ambiguity con- 

 cerning increment interpretation. To compare age 

 estimates obtained from different readers and to 

 quantify measurement error, every tenth specimen 

 in a tow subsample was aged independently by two 

 people. 



Preextrusion larvae from gestating females were 

 examined to characterize otolith growth prior to par- 

 turition. Samples for this study came from another 

 cruise of the RV David Starr Jordan ( 17-27 Febru- 

 ary 1993). Mature gestating female S.jordani were 

 collected by bottom trawl (26-m headrope, 15-cm 

 mesh, 1.25-cm codend), their ovaries were removed, 

 and embryos and prolarvae preserved in 95% etha- 

 nol. Prolarvae (i.e. hatched embryos) were staged in 

 the laboratory on the basis of a modified version of 

 Yamada and Kusakari's (1991) system and on the 

 basis of information on the temporal course of devel- 

 opment, inferred from laboratory experiments on 

 Sebastes flavidus (Eldridge 1 ). Specifically, the follow- 

 ing prolarval stages were assigned: 1 = parturition 

 in 5—6 d; 2 = parturition in 4-5 d; 3 = parturition in 

 3-4 d; 4 = parturition in 2-3 d; 5 = parturition in 1- 

 2 d; and 6 = parturition in 0-1 d. Parturition was 

 considered imminent in stage-6 prolarvae that had 

 depleted yolk reserves but had a well-developed oil 

 globule. After staging and measuring the prolarvae, 

 the otoliths were removed, affixed to glass slides, and 

 the otolith radius measured (^m). No more than five 

 prolarvae were examined from any single gestating 

 female. 



Otoliths of gestating and planktonic larvae (rc=60) 

 were examined with a scanning electron microscope 

 (SEM). To prepare otoliths for viewing, whole larval 

 specimens were dehydrated in successive rinses of 

 95 and 100% ethanol (24 hours in each) and im- 

 mersed in Spurr's embedding medium. Notochord 

 length (NL) was measured with the aid of a binocu- 

 lar microscope and calibrated ocular micrometer 

 scale. Each fish was covered with the medium and 

 oriented on its side in a silicone rubber embedding 

 mold (Haake et al., 1982). Curing required 24 hours 

 at 70"C. The otoliths from these larvae were hand 

 ground with 9-^m aluminum oxide microfinishing 



1 Eldndge, M. 1994. Tiburon Laboratory, Southwest Fisher- 

 ies Science Center, NOAA, 3150 Paradise Drive, Tiburon, CA 

 94920. Unpubl. data. 



film and mechanical lapping wheels to produce a mid- 

 sagittal section from one of the two sagittae. The 

 grinding progress was monitored by frequent inspec- 

 tion of the sample with both transmitted and inci- 

 dent illumination in a metallographic microscope. 

 Final polishing was achieved with 3- and 1-pm dia- 

 mond pastes. After the final polish, the mounts were 

 rinsed with distilled water and air dried. Immersion 

 and gentle agitation in a 5% solution of ethylene- 

 diaminetetraacetic acid (EDTA) (pH adjusted to 7 

 with potassium hydroxide [KOH] ) etched the otoliths 

 for observation under a SEM. Required etching times 

 of 10-40 seconds were determined by periodic rinses 

 and inspection with a light microscope. Owing to 

 shifting of the otoliths in the otic capsules or to twist- 

 ing or distortion of the larvae, approximately one 

 third of the samples resulted in a somewhat oblique 

 to transverse section through the sagitta. Measure- 

 ments of radii or increment widths were not signifi- 

 cantly affected by this occurrence because otoliths 

 have almost perfect radial symmetry during the ear- 

 liest postextrusion stages. The etched preparations 

 were then vacuum sputter-coated with gold palla- 

 dium before viewing with a back-scattered electron 

 detector (BED in the SEM. At least two micrographs 

 were taken for each sample, usually at 2,000— 3, 500x 

 and 10,000x. The photographic images were digitized 

 and measurements were made with the aid of an 

 image analysis software package (Optimus Corp. ). A 

 custom program was used for semi-automated mea- 

 surements of maximum otolith radius (from the cen- 

 ter of the otolith primordium) and maximum radius 

 to the presumptive extrusion check. The first 7-10 

 postextrusion increments were then marked and 

 measured. 



Results 



The distribution of S. jordani larvae in the survey 

 region was quite variable, with catch rates ranging 

 from to 10,670 larvae- 10 m~ 2 of sea surface. Like 

 many ichthyoplankton surveys, the distribution of 

 catch rates was strongly skewed. For example, the 

 five largest tows accounted for 3.33% of the effort, 

 yet produced 65.49^ of the catch. The overall mean 

 catch rate of larvae was 328 larvae-10 m 2 (rc = 150, 

 a =1,285, CV=392%). 



A total of 2,203 larvae was subsampled from the 

 total catch and aged by optical microscopy. Larvae 

 ranged in age from 0-34 d (median=4 d). Among the 

 aged larvae, the average radius of the otolith mea- 

 sured to the extrusion check (i? (J ) was 16.88 jim 

 (ct =1.41 iim), with 90% of all values lying in the 

 range 14.5-19.1 /jm (Fig. 2). 



