JOHNSON ET AL.: AGE. GROWTH. AND MORTALITY OF KING MACKEREL 



weight intervals (computed by the program) 

 were used. A maximum of 20 length-weight val- 

 ues was randomly selected for the analysis with- 

 in each qualifying length and weight interval. If 

 any length or weight interval contained fewer 

 than 20 values, all were utilized. 



Estimates of annual mortality rate A (after 

 Ricker 1975) were developed by catch-curve 

 analysis of south Florida length-frequency data. 

 These data were used because they best repre- 

 sented the king mackerel in U.S. waters accord- 

 ing to Trent et al. (1981). Since these data were 

 not separated by sex, two age-length keys were 

 developed, one combining males and females as- 

 suming a 1:1 sex ratio and the other assuming a 

 1 male:2 female ratio (the approximate ratio in 

 our collection). The length-frequency data were 

 converted to age-frequency distributions {N, = 

 number of fish caught in age-class i ) by applying 

 each of the combined age-length keys. Age classes 

 I through X of the resultant catch curves were 

 analyzed by 



1. Heincke's (1913) method; 



2. Jackson's (1939) method; 



3. Rounsefell and Everhart's (1953) method; 



4. Beverton and Holt's (1957) method, using 

 the mean of values computed with their 

 equation 13.4 between successive age 

 groups; 



5. Robson and Chapman's (1961) method, un- 

 corrected for possible age-length key bias; 

 and 



6. finding the slope (m) of a regression line 

 fitted to In (N,) and i and substituting in 

 the equation A — 1 — e m . 



RESULTS AND DISCUSSION 



Age 



The validity of using otoliths for estimating the 

 age and past growth history depends on these 

 structures being directly correlated with the 

 growth of the fish and on otolith mark formation 

 being periodic. We found the otolith radii to be 

 closely correlated to fork lengths, especially 

 when the data were transformed to represent a 

 "power" function. The "power curve" equation, 

 FL = 1.232 OR 1331 with correlation coefficient r 

 = 0.987, had a better fit than the linear equa- 

 tion, FL = 5.559 OR + 84.818 with r = 0.847. 

 This close correlation of OR and FL satisfied the 

 first criterion for validation of otoliths as an age 



determination structure. The second criterion, 

 mark formation of known periodicity, needed 

 further investigation. Beaumariage (1973) found 

 king mackerel with opaque margins during 8 mo 

 of the year (February-September); the highest 

 percentage of otoliths with opaque margins oc- 

 curred in May. He concluded, "Most otolith mar- 

 gins become opaque (form annuli) during April. 

 May, and June...." Fish in our collections exhib- 

 ited opaque margins in 11 moof the year with the 

 peak during May (54%); however, few fish were 

 collected during the winter months (November- 

 February). No month had a high percentage 

 (over 75%) of fish with opaque margins, and only 

 one month (March) lacked fish whose otoliths 

 had opaque margins (Table 1). 



In recent years the use of whole otoliths for 

 estimating the age of fish has been questioned. 

 Beamish ( 1979) indicated that a fish's age may be 

 underestimated using surface examination and 

 that otolith sections are more reliable. However, 

 we found 96.5% agreement between king mack- 

 erel age estimates (number of opaque marks) 

 comparing surface and sectional readings. This 

 indicates that our age estimations for whole oto- 

 liths are similar to those of sectioned ones. 



The agreement betw r een tw r o readers about the 

 number of marks on king mackerel otoliths was 

 98%. The number of otoliths found to be usable 

 was 1,449. 



Age and Size Composition 



Age composition of king mackerel varied 

 greatly among the areas (Table 2). Younger fish 

 were taken in northwest Florida, while older fish 

 were caught off Louisiana, particularly in 1978. 

 Fish of intermediate age were landed primarily 

 in Texas, South Carolina, and North Carolina. 

 The oldest females in our sample were 14+ yr 

 (over 1,400 mm FL) and the oldest males were 9+ 

 yr (970 mm FL;. 



Much age variation occurred within a single 

 length group in our data (Tables 3, 4) as it did in 

 Beaumariage's (1973) data. For example, we 

 found females 850-899 mm FL were 1-8 yr old 

 (Table 3). 



Back-Calculated Growth 



The weighted means of the back-calculated 

 fork lengths for male and female king mackerel 

 from all areas and years sampled in this study 

 are shown in Tables 5 and 6. Differences in mean 



99 



