to 1981; however, this was partly due to improvements in survey design 

 (NUSCo 1983). Female spawners and egg production peaked in 1982 with 

 the 1983 and 1981 estimates comparable in magnitude. 



Age and growth 



During 1983, the scales of 214 females and 188 males ranging from 

 44 to 465 mm were examined for age and measurements were made to each 

 annulus for the purposes of calculating growth. Some curvilinearity was 

 seen in a linear regression of scale and length, especially with larger 

 specimens, indicating probable heterogeneous growth of the scale and 

 fish (Fig. 5) . A non-linear length-scale relationship was used in the 

 back-calculation of length-at-age because it provided a better fit to 

 the data. Length at each annulus was calculated by the relationship: 



length = 3.557 (scale size)*^"^^^ for females, (n=216, r2=0.93) 



length = 3.777 (scale size)^"^^'^ for males, (n=193, r2=0.94) 



The mean calculated lengths-at-age (Tables 8 and 9) were larger than the 

 observed lengths except for ages 1 and 2 where resumption of seasonal 

 grovjth probably occurred before the scales were collected. The 

 calculated growth estimates were probably less reliable for older 

 specimens, particularly males, as calculated lengths became considerably 

 greater than observed lengths-at-age. A reverse Lee's phenomenon was 

 also observed where the calculated lengths of fish increased as age 

 increased. This may have been the result of size-selective mortality 

 that was greater on smaller fish of an age group (Tesch 1968) or due to 

 a bias in sample selection if only faster-growing larger specimens were 

 chosen for aging and scales from slower-growing fish were rejected as 

 unreadable. This may also have resulted in the apparent anomalous 

 increase in growth at age 9 in females and 8 in males. The smaller 

 number of older specimens examined also made their mean growth estimates 

 less reliable. 



22 



