302 



Fishery Bulletin 97(2), 1999 



<3> 



C 

 



C 



a 



200 300 400 500 600 700 800 



Lapillar area (^lm^) 



900 



10 



12 



14 16 1£ 



Radius of lapillus at hatching (nm) 



Figure 6 



Relation between the residuals from the mean SL and mean egg diam- 

 eter and (A) the lapillar area (LAi and (B) the lapillar radius at hatch- 

 ing (LRi for newly hatched cod larvae on the Scotian Shelf. Shown on 

 each plot are the predicted linear relationships and 95'"! confidence 

 intervals of the mean and of individual predictions. Regression rela- 

 tionships are; Residual = 0.00162 (±0.0003) x LA - 0.721 (±0.186), ;; = 

 37, ;•- = 0.38. P = 0.0001; Residua! = 0.126 (±0.03) x Lfl - 1.633 (±0.39), 

 n = 37. r- = 0.37, P = 0.0001. 



backcalculate to early periods of the life history. 

 Geffen (1995) commented on the variability in the 

 sizes of hatchmarks on the lapilli of cod larvae. She 

 reported estimate.s of otolith diameters at hatching 

 that varied by more than twofold ( 13.7-28.4 |im). In 

 Figure 8, we summarize published data on SL and 

 lapillar diameter at hatching for a range of cod .stocks. 

 The variability evident in Figure 8 suggests that the 

 choice of a "biological intercept" for cod is problem- 

 atical (Campana, 1990); a single "best" size at hatch- 



ing or otolith size at hatching clearly cannot be de- 

 fined for cod. However, the data do show that an over- 

 all relationship between the mean SL at hatching 

 and the mean diameter of lapillus exists. However, 

 given the limited data set, the relationship is statis- 

 tically insignificant (MDL = 1.73 xMSL -(-19.54, 

 /■'-=0.306, n=7, P>0.05). With these results, we rec- 

 ommend that extreme caution be exercised when 

 selecting parameter estimates for use in the back 

 calculation of size-at-age in cod. 



