558 



Fishery Bulletin 92(3). 1994 



were then dried, taped to an acetate sheet, inserted 

 between two other blank sheets, and pressed with a 

 Carver laboratory scale press for two minutes at 2,721 

 kg of pressure at 71°F. Because of the large size of weak- 

 fish scales, scale impressions were read with a stan- 

 dard microfiche reader at 20x. Those scales with po- 

 tential annuli crowded along the scale periphery were 

 also viewed at 48x under a dissecting microscope. Pre- 

 sumed annual marks were identified by standard cri- 

 teria (Bagenal and Tesch, 1978; Shepherd, 1988). 



One spiny ray from the dorsal fin and one soft ray 

 of the left pectoral fin were prepared from each fish. 

 Rays were serially sectioned by starting at their base 

 and cutting through most of their length at a thick- 

 ness of 400 urn with a Buehler low-speed Isomet saw. 

 Sections were then mounted on microscope slides 

 with Flo-Texx and read under a dissecting microscope 

 with transmitted light and dark field at 64x. Pre- 

 sumed annual marks were counted when they could 

 be identified as individual, opaque bands. 



Each hardpart was read twice by two separate 

 readers. Readings were done in a randomly selected 

 order, with no knowledge of collection date or fish 

 size. Hardparts were evaluated in terms of clarity of 

 presumed annual marks, ease of collection and pro- 

 cessing, and precision. Precision was measured by 

 average percent agreement within and between read- 

 ers, i.e. percent agreement within readers was cal- 

 culated for each reader separately and then averaged 

 for the two readers and percent agreement between 

 readers was calculated separately for each reading 

 and then averaged for the two readings. 



Validation of the otolith method 



Because otoliths were found best for ageing, additional 

 samples were collected for validation. During 1989- 

 92, 1,928 weakfish were collected from commercial 

 pound-net, haul-seine, and gill-net fisheries in Ches- 

 apeake Bay. During March-November when weakfish 

 are not present in the Chesapeake Bay, fish were col- 

 lected ( n =289 ) from the trawl fishery operating in North 

 Carolina shelf waters north of Cape Hatteras. 



The marginal increment method was used to vali- 

 date otolith annuli (Brothers, 1983; Casselman, 1987; 

 Hyndes et al., 1992). The translucent margin out- 

 side the proximal end of the last annulus was mea- 

 sured along the ventral side of the otolith sulcal 

 groove (Fig. 1). Measurements were taken with an 

 ocular micrometer to the nearest 0.038 mm (one mi- 

 crometer unit at a total magnification of 24x). 



Comparison of scales and otoliths 



To compare the otolith and scale methods in more 

 detail, 155 fish ranging from 140 to 845 mm TL were 



selected by stratified, random subsampling — strata 

 being otolith-determined ages — from a total of 300 

 fish collected in 1989 and 1992. Thirty fish were se- 

 lected from each of the age-strata, 1-4. Because older 

 fish were scarce, only 14 age-5, 16 age-6, two age-7, 

 two age-8, and one age- 10 fish were included. Al- 

 though most fish came from Chesapeake Bay com- 

 mercial fisheries, in order to increase the number of 

 older fish, 27 fish were collected in May 1992 at the 

 Delaware Bay Weakfish Sport Fishing Tournament. 

 We collected an additional 20 fish in August 1992 to 

 include fish from each of the summer months for mar- 

 ginal increment and back-calculation analyses. 



Hardparts were prepared as described for the pre- 

 liminary comparison and read twice by each of two 

 readers. An effort was made to determine annuli on 

 scales based only on physical criteria and not to as- 

 sign annuli based on any preconceived ideas of 

 growth (Casselman, 1983). Reading order was ran- 

 domized and collection date and fish size were un- 

 known. Each reader recorded the number of pre- 

 sumed annuli and a "+" if there was growth beyond 

 the last annulus or a "*" if the last presumed annu- 

 lus was forming or had just formed (Casselman, 

 1987 ). After all hardparts had been read, we assigned 

 ages using a January 1 birthdate, knowledge of the 

 time of annulus formation, the relative growth of the 

 hardpart margin, and date of capture. 



Variability within reader, between readers, and 

 between hardparts was analyzed by percent agree- 

 ment. When an individual reader's counts of pre- 

 sumed annuli disagreed, a third reading was made. 

 When readers' ages disagreed, a third reading with 

 both readers present was made to resolve the dis- 

 agreement. 



To compare time of annulus formation and its vari- 

 ability in scales and otoliths, mean monthly relative 

 marginal increments and their ranges were calcu- 

 lated and plotted (April-October). Relative marginal 

 increments were calculated by dividing the marginal 

 increment by the hardpart radius. All ages were 

 pooled. Additionally, those hardparts which had been 

 designated as having an annulus on the margin ("*") 

 were reviewed and their time of collection recorded. 



To determine marginal increments and to conduct 

 back-calculation analyses, hardparts were measured 

 by using a Via 100 camera/monitor system with a 

 dissecting microscope at 24x. Otolith radius ( OR) and 

 otolith annular radius (OAR) — the distance from the 

 nucleus to the proximal edge of each annulus — were 

 measured along the ventral arm of the sulcal groove. 

 Scale radius (SR) and scale annular radius (SAR) 

 were measured along the left radius (Ricker, 1992). 

 Marginal growth was measured from outside the last 

 annulus to the hardpart edge. 



