mackerel was measured to the nearest millime- 

 ter. Sagittal otoliths were removed, washed, and 

 stored dry. The clearest, most legible otolith from 

 each fish (based on visual observation) was exam- 

 ined to estimate age and growth. 



Whole otoliths were placed in a black-bottomed 

 watch glass containing 100% glycerin and exam- 

 ined with a binocular microscope at 28 x using 

 reflected light. Otolith radius (OR) was measured 

 in ocular micrometer units (1 unit = 0.0363 mm) 

 on the posterior surface from the focus to the dis- 

 tal margin along the axis of the sulcus acousticus 

 (Powell 1975). Growth marks were counted and 

 measured from the focus along the radius to their 

 distal edge. The marks were opaque (light) under 

 reflected light, while the interspaces were hya- 

 line or translucent (dark). 



Otoliths were classified into age groups based 

 on the number of opaque nonmarginal marks 

 (Powell 1975). A mark was considered complete 

 when a hyaline (dark) interspace or margin was 

 visible from successive growth. Three readers in- 

 dependently examined each of 520 otoliths to test 

 the precision of our ageing technique. This infor- 

 mation was analyzed using the method of 

 Beamish and Fournier (1981). All other otoliths 

 were independently examined by two readers; if 

 their results did not agree, the data were not 

 used. 



To compare age estimates based on surface 

 (whole) and internal (sectional) examination, we 

 sectioned 70 otoliths which had been previously 

 examined on the surface (2-10 otoliths from dif- 

 ferent fish from each age 0+ through 8-1- ), follow- 

 ing the methods of Johnson et al. (1983). 



We determined time of annulus formation and 

 validated our ageing technique by comparing 

 monthly percentage frequencies of otoliths with 

 opaque margins. A high percentage frequency 

 (>45%) indicated recent annulus formation. We 

 used a chi-square test to compare the monthly 

 frequencies. 



The relationship between otolith radius and 

 fork length was determined and used to back cal- 

 culate fork lengths at earlier ages (Tesch 1971; 

 Ricker 1975; Everhart et al. 1975). We used anal- 

 ysis of covariance (ANCOVA) with age as the co- 

 variate to test for differences in growth rates 

 (lengths at age) of fish collected in different loca- 

 tions, by different gears, and of different sexes. 

 Mean back-calculated lengths were used to calcu- 

 late von Bertalanffy (1938) growth parameters, 

 employing a computer program developed by 

 Abramson (1971). 



FISHERY BULLETIN: VOL. 85, NO. 4 



RESULTS AND DISCUSSION 



Validation 



Age validation has often been overlooked in the 

 age and growth literature (Beamish and McFar- 

 lane 1983). Although there are numerous meth- 

 ods available to establish the annual nature of 

 otolith growth rings, we applied marginal incre- 

 ment analysis, because it was the only practical 

 method to use on this migratory, pelagic species. 



Annulus formation occurred in March, April, or 

 May (Fig. 1). A chi-square test (x^ = 338.47, 

 df = 1, P < 0.001) showed a highly significant dif- 

 ference between the occurrence of otoliths with 

 opaque margins in these months versus the other 

 nine months of the year. Our findings are in 

 agreement with Powell (1975) in that the main 

 period of opaque mark formation was in the 

 spring or early summer. He reported mark forma- 

 tion in May, June, and July by examination of 

 marginal increments. Previously Klima (1959) 

 described both summer and winter growth rings 

 and evaluated the marginal condition to decide 

 that marks were deposited annually. Our obser- 

 vations on the appearance of annuli in Spanish 

 mackerel otoliths agreed with Powell (1975), in 

 that we also were unable to discern the "first win- 

 ter mark" that Klima (1959) described. 



Age 



To estimate the precision of our ageing, we 

 compared sections to whole otoliths and evalua- 

 tions by different readers. Examination of 70 sec- 

 tioned otoliths provided a 97.4% agreement with 

 previous surface examination of the same 

 otoliths. Surface age determinations of three 

 readers on 520 otoliths had a 97.7% agreement. 

 Using the technique of Beamish and Fournier 

 (1981), the index of average percent error was 

 0.3273, which we think is excellent. 



Of 1,929 Spanish mackerel examined, 1,787 

 (92.6%), ranging from 148 to 802 mm FL, were 

 aged. The oldest female was 9 years old, while the 

 oldest male was 7 years old. Powell's (1975) oldest 

 fish, a female, was 8 years old, while Klima's 

 (1959) oldest males and females were both 6 years 

 old. These data and the data presented in Tables 

 2 and 3 indicate that females live longer than 

 males. 



We found a wide range of lengths within an age 

 group for both sexes (Tables 2, 3), as did Powell, 

 with some Spanish mackerel of age through 5 in 



778 



