Lowerre-Barbien et al.: A comparison of otolith and scale ageing methods for Cynoscion regalis 



561 



fish's life history. Additionally, of the 2,217 otoliths 

 examined (ages 1-10), all those in the process of form- 

 ing or which had just formed annuli were collected 

 in March-May. Thus, we assumed for ages 1-10 that 

 the otolith method provided accurate ages. 



Comparison of scales and otoliths 



Scales were consistently more difficult to read than 

 otoliths, and confidence in scale readings was often 

 low. Percent agreement within and between readers 

 was fairly consistent for both hardparts. However, 

 otoliths showed much higher agreement (98-100%), 

 than did scales (78-80%) (Table 2). Although agree- 

 ment between scales and otoliths was fairly high, 

 79%, agreement decreased with increasing age. Of 

 32 disagreements, only 6 differed by more than one 

 year (Fig. 5). However, 4 of the 5 fish older than age 

 6 were underaged by scales and two of the oldest fish, 

 age 10 and 8, were underaged by 3 years. Scales from 

 older fish, if they showed more than 6 annuli, had 

 marks which were severely crowded and fragmented 

 even when viewed at higher magnification (Fig. 6A), 

 whereas otoliths from these same fish showed clear 

 annuli (Fig. 6B). 



Although the number offish underaged was small, 

 their effect on estimating growth curves would be 

 dramatic. Mean body size at age based on scales, al- 

 though slightly curvilinear, showed no clear indica- 

 tion of an asymptote (Fig. 7A) and thus would not be 

 appropriate for fitting a von Bertalanffy growth curve 

 (Gallucci and Quinn, 1979). In contrast, mean body 

 size at age based on otoliths showed the clear begin- 

 nings of an asymptote (Fig. 7B). 



Although sex of the fish had no effect on the preci- 

 sion or repeatability of scale readings, it did affect 

 accuracy. Agreement of scale ages among and be- 

 tween readers was quite similar when calculated 

 separately by sex, ranging from 75 to 79.5%. How- 

 ever, agreement between scale and otolith ages, or 

 accuracy, was significantly different for males and 

 females <x 2 =6.25, ra=154, P<0.05). Of the 32 discrep- 

 ancies between scale and otolith ages, 26 of them 

 were males. Even if the fish greater than age 6 are 



discounted, there is still a significant difference 

 (X 2 =5.79, n = 149,P<0.05>. 



Time of annulus formation is not the same for 

 scales and otoliths. Both hardparts showed only one 

 trough in their mean monthly marginal increments 

 (Fig. 8). However, otoliths with annuli on their mar- 

 gins were collected only during a discrete time pe- 

 riod, 1 April-1 June, while scales in the process of 

 forming annuli were collected from mid-April to mid- 

 August, although most scales formed annuli in Au- 

 gust. The variable and extended time of scale annu- 

 lus formation is represented by the shallow trough 

 (Brothers, 1983) and the larger standard errors of 

 the scale marginal increment plot, as compared with 

 that of otoliths (Fig. 8). 



Although total length on hardpart size relation- 

 ships for both scales and otoliths showed linear 

 trends (r 2 =0.94 and 0.88 respectively, « = 175, 

 P=0.0001), the total length on otolith relationship 

 showed seasonal variation. When a single age class 

 (age 3) was marked by season of collection and plot- 

 ted against the linear relationship predicted by the 

 total sample (Fig. 9), all fish collected in April and 

 May had smaller than predicted otolith radii, 

 whereas fish collected in August and September had 

 larger than predicted radii. Fish collected in June 

 and July were intermediate, although most of their 

 radii were also smaller than predicted. Scales from 

 the same fish did not show similar seasonal trends. 



Back-calculation equations of total length on hard- 

 part size were calculated only for fish collected at 

 the beginning of the growing season, in April and 



10n 



CD 

 CD 



ro 



0) 

 CO 



o 



CO 



Otolith age 



Figure 5 



Weakfish, Cynoscion regalis, assigned ages from 

 scales and otoliths. The number of fish each point 

 represents is indicated. The 45 line represents 100 1 ?* 

 agreement. 



